abstract

Microseismic monitoring (MSM) of hydraulic fracture treatments is routine in North America and has added significantly to our understanding of fracture growth. The interpretation of microseismic images is advancing steadily, extracting more information from event patterns, temporal evolution, and acoustic waveforms. The increasing amount of information from MSM provides significant opportunities to improve stimulation designs, completion strategies, and field development. However, the applications of microseismic interpretations are many times ill-defined, overlooked, or not applied properly. Numerous applications of microseismic measurements have been documented in technical publications, typically in the form of case histories focused on specific applications. The industry has lacked a compilation and comprehensive discussion of microseismic applications. This paper presents a practical guide for the engineering application of microseismic interpretations, documenting reliable application workflows while highlighting the consequences of misapplication of microseismic interpretations.

The application of MSM starts with a reliable interpretation of fracture geometry and complexity, but the real value is in the application of the interpretation. This paper divides microseismic applications into three categories, real-time, completion strategies & stimulation design, and field development. The MSM interpretation requirements for each category are documented and a comprehensive guide to properly applying these interpretations is presented. Applications issues such as determining the "effective" fracture surface area, the relationship between microseismic behavior and well performance, and fracture model calibration are addressed.

There is a growing interest in advanced processing such as moment tensor inversion (MTI) and b-values to determine focal mechanisms, source parameters, and failure mechanisms associated with the microseismic events. However, the engineering application of these interpretations is not well understood. This paper includes a discussion of the applications of advanced processing results, emphasizing how the limitations and uncertainties of the processing affect the subsequent applications.

Introduction

Microseismic mapping of hydraulic fracture treatments is now commonplace in North America and has played an important role in the advancement of stimulation and completion techniques, especially for unconventional reservoirs (King et al., 2008; King, 2010; Waters et al., 2009a; Cipolla et al., 2010, 2011a). The initial focus for microseismic technology was improving the acquisition of the seismic waveforms, more accurate event locations (processing), and visualizing the images with the relevant geological and geophysical context (Fig. 1). However, current efforts focus on more value-added applications of microseismic data, including real-time treatment control, integration with geology, geophysics, geomechanics, and fracture modeling, and utilizing microseismic data in reservoir engineering workflows (Fig. 1). These value-added applications of microseismic data are the primary focus of this paper, as they have the potential to add significant insights into hydraulic fracture "effectiveness" and completion "efficiency" that could result in considerable improvements in well performance.

Fig. 1

Initial focus of microseismic technology was improving acquisition, more reliable event locations, visualization of images with geological and geophysical context, and interpretation of the images. Current efforts now include more focus on value-added applications, including real-time placement, integration with geology, geophysics, geomechanics, and fracture modeling, and using microseismic images in reservoir engineering workflows.

Fig. 1

Initial focus of microseismic technology was improving acquisition, more reliable event locations, visualization of images with geological and geophysical context, and interpretation of the images. Current efforts now include more focus on value-added applications, including real-time placement, integration with geology, geophysics, geomechanics, and fracture modeling, and using microseismic images in reservoir engineering workflows.

The application of microseismic measurements requires a reliable and detailed interpretation of the geophysical images and data. Cipolla et al. [2011b] present a guide for interpreting microseismic measurements, which serves as a starting point for this paper. Since the interpretation and application of microseismic measurements are closely linked, the interpretation guidelines from this previous work are reviewed to provide the necessary background for the subsequent applications.

Interpretation Guidelines (adapted from SPE 144067, Cipolla et al., 2011b)

The typical interpretation of microseismic event patterns consists of six primary geometrical observations and the estimation of stimulated volume (SV):

  1. Fracture Length

  2. Fracture Height

  3. Fracture Azimuth

  4. Fracture Complexity (i.e. – network or planar fractures)

  5. Fracture Location with respect to the perforations, frac port, or other exit point from the wellbore.

  6. Anomalous behavior (i.e. – fault activation, asymmetry, etc.)

  7. Stimulated Volume - SV (SRV, ESV, etc.)

Basic Microseismic Location Interpretation

Reliably interpreting microseismic locations requires an understanding of the inherent uncertainty in the measurements, the mechanisms that generate microseisms, the reliability of input data, and the accuracy of the processing algorithms. The interpretation of microseismic data begins with the fundamental source characteristics including event locations and times, magnitude and P/S amplitude ratios along with QC attributes consisting of at least signal-to-noise-ratio (SNR) and location error ellipsoids. It is essential to account for observation well distance and location bias because microseismic images are often misinterpreted without these corrections, potentially leading to erroneous applications. The final interpretation requires integration with the geological setting, visualization of the events in space and time, and simple fracture mechanics that link the microseismic image to the treatment data. The basic interpretation workflow and guidelines are summarized below.

  1. Uncertainty. The first step in the interpretation of microseismic measurements is to evaluate the uncertainty in the event locations. The location uncertainty for most events will differ in different directions (depth, distance, and azimuth). SNR can be used as an indicator of event waveform quality, with higher SNR typically associated with more accurate event locations.

    • Filtering.The microseismic events should be filtered using various SNR cutoffs and the resultant event patterns and error ellipsoids compared. Significant changes in event patterns and/or increases in location uncertainty could indicate that events with higher location uncertainty are introducing errors into the interpretation. Individual data sets include distributions of SNR and uncertainty: ranging from events with low to high confidence. The data should be filtered to minimize events with high location uncertainty prior to the interpretation, while including sufficient events to characterize the geometry without imposing location biases. The appropriate SNR and error cutoff will depend on the overall quality and number of events.

    • Velocity Model.Uncertainty in the velocity model is often overlooked when interpreting microseismic event patterns. In some cases, errors in the velocity model may result in significant interpretation errors. The accuracy of the velocity model can be verified by the accuracy of perforation shots if a clear s-wave is observed. The geophysical work-product may include velocity model uncertainty in the calculation of event location uncertainty (e.g. error ellipsoids). Quality control during the geophysical processing is a critical step prior to interpretation to ensure that the microseismic locations are correct and that the uncertainty and confidence in the results are communicated to the interpreter to avoid over interpretation of the microseismic data.

  2. Observation Well Bias.The location of the observation well with respect to the treatment well and the location of microseismic activity can introduce errors into the interpretation. Small events can only be detected if they are close to the sensor array and at some distance even the largest events cannot be detected. The strength of the microseismic events can vary considerably, resulting in an apparently greater event density close to the sensor array. Increased background noise levels can also impact detectability. Magnitude-distance plots provide QC of the observation well distance bias. Directional biases can also occur in cases where the events have similar failure mechanisms, such that the source radiation results in nodal planes where either the p- or s-waves are not observed. Microseismic detection algorithms that rely on indentifying both wave types will then result in fewer events in these nodal directions. P/S amplitude ratio plots provide QC of potential directional biases. Understanding the detection limit is important when interpreting fracture geometry and comparing stimulation treatments. In cases where the interpretation indicates that the entire fracture geometry was not detected due to biases, the geometry must be assumed to be symmetrical.

  3. Geologic Environment. The interpretation of microseismic event patterns requires a basic knowledge of the geologic environment in which the hydraulic fracture propagated. Event patterns can vary significantly depending on the reservoir fluids, stress regime, existence of natural fractures, matrix permeability, and rock properties. An understanding of the geologic environment and the basic mechanisms that generate microseisms will help constrain the interpretation. Note that the interpretation should identify anomalous microseismic events that are likely not associated with fracture propagation, such as stress-induced fault activation.

  4. Visualization and Integration.The bulk of routine microseismic interpretation focuses on integration of the microseismic data with geology, seismic, and treatment data and visualization of the event patterns. The visualizations can be two dimensional or three dimensional and include both geologic and seismic information and include Temporal and Spatial Evaluation, Stimulated Volume (SV), Spatial and Temporal Event Histograms and Anomaly Identification. These interpretation tools are described in detail by Cipolla et al. [2011b] and will be illustrated later in the text using examples and case histories.

  5. Mass Balance and Simple Fracture Mechanics Integrating the volume of fluid, proppant pumped and the net pressure at the end of the treatment with the microseismic event pattern can help constrain the interpretation. Simple fracture mechanics equations can be used to calculate the average fracture width, total fracture area, and total fracture length from the microseismic fracture height, estimated net pressure and fluid efficiency. In the case of complex fracture networks, total fracture length can be combined with stimulated volume calculations such as ESV or SRV to estimate network fracture spacing.

With a reliable interpretation of the microseismic image providing fracture geometry, azimuth, location, and complexity along with the identification of anomalous behavior (if present), the microseismic application workflows can be executed. The mechanism of the microseismic data in relation to the hydraulic fracture will be discussed later, but basic interpretation makes the inherent assumption that the hydraulic fracture(s) is proximal to the microseismic events. Although there are very limited direct observations, extensive work was performed at the M-Site (MWX project) to validate the application of microseismic mapping to image hydraulic fracture geometry (Warpinski et al., 1998), while cored hydraulic fractures at the Mounds drill cuttings injection experiments confirmed microseismic fracture geometry (Moschovidis et al., 2000). In addition, observations of offset wells being contacted (i.e. – "killed") by fracturing fluid during an offset well treatment also support microseismic interpretations of fracture geometry (Fisher et al., 2002). Note that the interpretation should identify anomalous microseismic events that are probably not associated with fracture propagation (Cipolla et al, 2011b).

Advanced Processing, Stimulated Volume, and Fracture Area

Beyond the microseismic locations, additional parameters can be used to characterize the deformation associated with the microseismic locations. It is important to understand the value and limitation of advanced seismic processing and interpretations to properly apply the results. The following sections discuss the issues associated with applying advanced microseismic processing and stimulated volume interpretations.

Advanced Processing and Fracture Area

In addition to event location and magnitude, there are ongoing efforts to extract additional information from the seismic waveforms. Advanced processing may provide important insights into the details of the failure mechanisms that can help define natural fracture characteristics and constrain hydraulic fracture models. However, these advanced processing techniques are relatively new and untested, and the limitations are not well understood by engineers when applying the results. It is important to understand the relationship between basic and advanced processing results, actual fracture geometry, and well productivity.

Microseismic activity rates.An obvious metric of a microseismic data set is the number of events. However, the number of events should not be misinterpreted as the true activity rate. A better measurement of activity is the total source strength or cumulative seismic moment (moment magnitude is the logarithm of this value), as discussed by Maxwell et al.,(2006). In terms of the geomechanical deformation associated with each microseism, a relatively large magnitude event can represent the same total rock strain as numerous smaller magnitude events. Small variations in detection distance and minimum SNR can result in big differences in the number of events. Cumulative moment is less sensitive to these variations, because most of the moment is associated with the bigger events. Applications comparing microseismic data often use activity rates as one of the comparisons; the ‘Comparative Interpretation’ section of this paper reviews the proper workflow to enable such a comparison.

Frequency-magnitude relationships. Analysis of the relationship between frequency (or number: N) and magnitude (M) generally follows a Richter-Gutenberg power law relationship of the form: log N = a - b M. The so called "b-value" or slope of the relationship is a good indicator of fault activation. For most fracs the b-value is approximately 2 (meaning there are 100 times more events for a decrease of one magnitude unit) while faults result in a b-value of approximately 1 (Maxwell et al., 2009). Geo-hazard assessment using this observation is described later in the paper. Normalizing the frequency- magnitude for uniform detection is also an important aspect of applications requiring a comparison of activity rates as described later.

Microseismic source mechanisms.The radiation pattern or directionality of p- and s-waves amplitudes can be used to define the failure mechanisms: the orientation of the failure plane and type of failure (e.g. shear or tensile). The simplest form is to assume shear failure and use the corresponding radiation characteristics to define the orientation of the failure plane. This can be graphically represented by a so-called "beach-ball diagram" akin to earthquake seismology. Alternatively, a technique called "moment tensor inversion" can be used to estimate the fracture plane and the type of failure simultaneously. To determine the mechanism for a single event, the radiation pattern must be determined in different directions either through surface or multiple downhole monitoring. Accuracy or confidence in the resulting moment tensor is critical, with recent research highlighting potential significant errors in both the failure plane orientation and the type of failure (e.g., Kim, 2011). Moment tensor can be plotted using standard seismology plots of radiation patterns or source plots, although new visualization is becoming available for engineering applications with more intuitive depiction of fracture planes and displacement vectors (Leaney and Chapman, 2011). Moment tensor is potentially of interest, enabling shear failure to be distinguished from tensile failure as well as identifying fracture opening or closing. However, the geomechanical interpretation of moment tensor data remains uncertain at this point in time: since the seismic energy associated with either fracture opening or closing will be relatively small in comparison to shearing because most rocks tend to be relatively weak in tension. Maxwell and Cipolla (2011c) argue that the majority of microseismic deformation tends to be in shear, and that a geomechanical based fracture model is needed to reconcile the paradox between shear microseismic deformation and tensile fracture opening and closing. This will be further discussed in the "Misinterpretation" section. A final category of mechanism analysis is more appropriate for single downhole monitoring, where groups of events are used to define a common failure mechanism. One form of this is plotting p- to s- amplitude ratios as a function of azimuth and fitting a theoretical radiation pattern (e.g. Rutledge et al.). The main applications of mechanisms are to identify fault activation or examine potential fracture complexity from interaction with pre-existing fractures.

Microseismic source attributes.Source parameters define the geomechanical strain occurring at the microseismic source. The most common is the seismic moment (defined as shear modulus*area*displacement) which is used to compute the moment magnitude (Mw = 2/3 log Mo -6) discussed above. The moment is potentially of interest since the product of area and displacement defines a fracture volume associated with the microseismic event. To explore this microseismic fracture volume, we assume that the displacement is a tensile fracture opening and hence the microseismic volume represents the fracture storage capacity. (If the displacement is a shear slip rather than a tensile opening, then the volume available for storage will be lower, including only the volume opened by misalignment of the irregular surfaces after the shear event.) The fracture surface area is also of interest. The dominant frequency can be used to estimate a slip area, assuming some dynamic failure model. Additional parameters related to stress release and seismic energy can also be considered. In terms of the microseismic fracture volume and area, it is insightful to compare the estimated values with independent estimates of a complex fracture model to the microseismic image [Cipolla et al., 2011a;,Weng et al., 2011]. The microseismic area is also much smaller than the modeled fracture area.

Table 1

Comparison of Microseismic Area, Volume, and Energy with Hydraulic Fracture Treatment Energy and Geometry

Stage 1Stage 2Stage 3Stage 4
MS Volume (bbl) 0.82 0.15 0.075 0.30 
Frac Volume (bbl) 25300 25300 25300 25300 
0.0033% 0.00059% 0.00030% 0.0012% 
MS Area (ft2) 19375 4018 3533 9040 
Frac Area (ft2) 7000000 6200000 
0.28% 0.057% 
MS Energy (J) 62500 11300 5700 22950 
Frac Energy (J) 149040000000 149040000000 149040000000 149040000000 
0.0042% 0.00076% 0.00038% 0.0015% 
Stage 1Stage 2Stage 3Stage 4
MS Volume (bbl) 0.82 0.15 0.075 0.30 
Frac Volume (bbl) 25300 25300 25300 25300 
0.0033% 0.00059% 0.00030% 0.0012% 
MS Area (ft2) 19375 4018 3533 9040 
Frac Area (ft2) 7000000 6200000 
0.28% 0.057% 
MS Energy (J) 62500 11300 5700 22950 
Frac Energy (J) 149040000000 149040000000 149040000000 149040000000 
0.0042% 0.00076% 0.00038% 0.0015% 

As noted above, the volume associated with the microseismic events is an extremely small fraction of the injected volume. Mass conservation, along with the comparably small energy balance percentages, suggests that the majority of the hydraulic fracture dilation is aseismic (i.e., not detected seismically). Maxwell and Cipolla [2011c] describe the aseismic deformation as low frequency, outside the bandwidth of seismic instrumentation. Furthermore the microseismic deformation is likely shear so that in reality only a small portion of the total microseismic deformation is fracture dilation. It should also be noted that the microseismic volume is an extremely small percentage (roughly 0.0000001) of the stimulated volume of about 10,000,000 m3. If this deformation is assumed to represent the stimulated volume fracture porosity, the extremely low value of this ratio casts doubt on what is sometimes described as the frac "rubblizing" the formation. The microseismic fracture area is a more significant proportion (but still less than 1%) of the hydraulic fracture area, which supports the fact that the microseismic activity describes the fracture geometry but not the entire fracture deformation. Regardless, the relative microseismic deformation is useful for applications involving calibrating a fracture/geomechanical model.

Stimulated Volume and Well Productivity

Microseismic images can be used to estimate the portion of the reservoir that was affected by the hydraulic fracture treatment, often referred to as Estimated Stimulated Volume (ESV) or Stimulated Reservoir Volume (SRV), but denoted stimulated volume (SV) in this text. SV is calculated based on the spatial distribution and density of microseismic events. These calculations can be performed using various 3D algorithms that vary in their approach to interpreting the microseismic image (Daniels et al, 2007;Mayerhofer et al., 2008) and it may not be possible to directly compare stimulated volume calculations from different algorithms. In some cases well productivity can be correlated to SV (Fig. 2), but these correlations are typically limited to specific regions within a field and similar stimulation designs (e.g. – water-fracs with similar volume and size of proppant). SV can also be correlated to production profiles in horizontal wells (Fig. 3). However, SV cannot provide insights into the actual fracture structure and is often insufficient when treatment designs and/or geologic environment vary significantly.

Fig. 2

Correlation between SV and well productivity in the Barnett Shale (From SPE 90051; Fisher et al., 2004).

Fig. 2

Correlation between SV and well productivity in the Barnett Shale (From SPE 90051; Fisher et al., 2004).

Fig. 3

Example of correlating ESV to stage productivity in the Eagle Ford (From SPE 136873; Inamdar et al., 2010).

Fig. 3

Example of correlating ESV to stage productivity in the Eagle Ford (From SPE 136873; Inamdar et al., 2010).

Stimulated volume calculations may provide an estimate of the maximum extent of the hydraulic fracture. However, numerous corrections are required before the most likely location of the hydraulic fracture can be determined. The application of stimulated volume calculations is similar to peeling an onion, it requires peeling away a number of layers before the useful portion is discovered. The corrections include location uncertainty and eliminating stress induced events (Cipolla et al., 2011b). Once the most likely location of the hydraulic fracture is determined, the next step is to evaluate fracture complexity. In some cases where fracture growth is primarily planar, stimulated volume is misapplied, as SV is only an appropriate measure for complex hydraulic fractures. Cipolla et al. [2008a] introduced the fracture complexity index (FCI), which provides a method to estimate fracture complexity using the ratio of the microseismic image width and length. Note that the microseismic width should be corrected for location uncertainty to avoid misinterpreting location uncertainty as an indication of fracture complexity. When FCI is less than 0.25, planar fracture growth should be considered for subsequent applications.

When the microseismic interpretation indicates that hydraulic fracture growth is complex, the most important parameters required for subsequent applications are hydraulic fracture geometry and the distribution of conductivity. Unfortunately, stimulated volume calculation, even after the above corrections, cannot provide insights into the underlying hydraulic fracture structure (i.e. – geometry) or the location of proppant (i.e. – distribution of fracture conductivity). In addition, although SV may correlate to hydrocarbon production in some cases, geomechanical effects such as closing of propped and partially propped fractures and/or decreasing propped fracture conductivity due to increasing closure stress may render the correlations invalid as drawdown increases with time in the fracture system.

Stimulated volume calculations provide an important quantitative interpretation, but the applications are more qualitative. In the past, SV was the primary interpretation for comparing microseismic images, but with the introduction of complex hydraulic fracture models and "microseismic to reservoir simulation" workflows there are now more rigorous methods for applying microseismic measurements. Stimulated volume may continue to be an important interpretation. However, future applications of SV will probably focus on fracture model calibration, evolving into an interpretation of "large scale" fracture location rather than a region of enhanced permeability.

Comparing Microseismic Images

Many applications of microseismic interpretations require comparing microseismic images to identify the impact of changes in either fracture design or completion methods, or to assess the impact of changes in geology on the microseismic response. This section provides some guidelines to apply whenever the relative changes or differences between two microseismic results are significant. The workflow associated with comparing the relative microseismic responses between fracs (stages or wells) is somewhat simpler than the interpretation of the absolute fracture geometry discussed later. Comparison of the relative microseismic responses typically involves assessing if the fracture geometry or induced microseismic deformation differs. Relative interpretation of the fracture geometry is easier than the complete absolute interpretation of geometry, for example when two fracture images are close together and tend to have similar accuracies (i.e. - potential systematic location errors associated with the velocity model are the same in both cases) and only the precision associated with input arrival times and directions need to be considered. Comparison of the microseismic deformation can be assessed by the number of microseismic events, but more robustly by the total strength of the microseisms (cumulative moment), as discussed previously. Diligent filtering of the data for consistent geophysical sensitivity and confidence between fracs is required to ensure appropriate comparison. Both aspects of comparative interpretation are highlighted here using data examples.

Fig. 4 shows a map view of two stages in a horizontal treatment well in the Barnett Shale. One stage is approximately 2000’ from the monitoring well and the other about 2400’. Microseismic locations along with an interpreted geometry encompassing 95% of the data are also shown. The event locations have been normalized using a minimum SNR cut-off of 5 for both stages. Note that the more distant stage is slightly larger in length (transverse to the well), and that there is an overlap of almost 300’ between the two stages along the well. The images of these two stages are now compared to interpret the differences in geometry imaged with the microseismicity.

Fig. 4

Map view of two stages in a horizontal treatment well. Top panel perf locations and the interpreted extent of the microseismicity. Middle panel showing microseismic events and the location uncertainty associated with the direction to the microseismicity. Bottom panel shows the location uncertainty associated with the horizontal offset direction.

Fig. 4

Map view of two stages in a horizontal treatment well. Top panel perf locations and the interpreted extent of the microseismicity. Middle panel showing microseismic events and the location uncertainty associated with the direction to the microseismicity. Bottom panel shows the location uncertainty associated with the horizontal offset direction.

First, the two stages are overlain with uncertainties in order to examine whether the slightly greater transverse length is significant. In this geometry, the location error in the transverse component of the frac is dominated by the geometric directional error, determined to be 4° in this case. This results in a ‘cone’ shape increasing with distance away from the monitoring well, as shown in the middle panel. The interpreted lengths both fall within these average uncertainty lengths, indicating that the interpreted extremes are consistent within location uncertainty, and therefore no significant differences in lengths are found between the two stages. More statistically sophisticated comparisons can be made, including statistical tests between the interpreted geometries. The final panel of Fig. 4 shows the overlap in the stages with the average location uncertainty in the offset direction, as a function of radial orientation from the monitoring well. This offset uncertainty is controlled by the relative timing accuracy between the s- and p-waves, and is determined to be about 50’. Therefore the observed overlap between stages is significant, and can be used to assess the completion strategy, such as making trade-offs between favorable fracture network connection versus the consequences of closely spaced stages resulting in uneconomic overstimulation or the closure of fractures stimulated in earlier stages.

The second aspect of a comparative interpretation involves comparison of the number of events detected between two fracs. Fig. 5 shows a plot of magnitude versus distance for two stages. More events are detected during the closer stage B than the more distant stage A, although the smallest magnitude event that is detected is lower. The impact of both aspects is needed to determine if one frac is inherently more microseismically active or if the interpretation is affected by observation well bias. Table 1 shows the number of events recorded for both stages and the corresponding cumulative seismic moment. As discussed previously, cumulative moment is the more robust metric for microseismicity rate or deformation comparison, since one relatively large magnitude microseismic event can be the deformation equivalent of numerous small magnitude events.

Fig. 5

Magnitude versus distance comparison of two stages. The red line represents the detection limit of the smallest event detected at various distances. The horizontal blue line represents the minimum magnitude that is "uniformly" detected at all distances (-1.8). The green line represents the "magnitude of completeness": the value above which all events have been recorded (-1.7).

Fig. 5

Magnitude versus distance comparison of two stages. The red line represents the detection limit of the smallest event detected at various distances. The horizontal blue line represents the minimum magnitude that is "uniformly" detected at all distances (-1.8). The green line represents the "magnitude of completeness": the value above which all events have been recorded (-1.7).

In this example, a higher minimum SNR of 3 was used for event detection for the further stage, compared to the 2.5 used for the closer stage. In order to compare these stages, a consistent detection threshold must be used. The first step is thus to filter both stages and consider only events with a SNR of 3 or greater. The corresponding number and cumulative moment are also shown in Table 2. In this case the results do not change significantly and stage B still has more events and a lower total strength. To fully quantify the microseismic activity in each stage, magnitude- frequency analysis is required. Fig. 6 shows the comparative magnitude and cumulative frequency data for this example. Note that at low magnitudes, the curve becomes flatter corresponding with the inability to detect small magnitude events. Above magnitude -1.7, all events in both stages are detected. Note that the frequency-magnitude plot shows a linear power law relationship above this level as expected. This "magnitude of completeness" (defined as the magnitude above which all events have been detected) is slightly higher than the minimum detectable magnitude at the largest offset, corresponding to a value of about -1.8. The total strength of events above this cut off is also included in Table 2 as well as shown graphically in Fig. 6. It is interesting to note that when the data are normalized, stage A has significantly more events and much greater total strength than stage B. Applications involving comparison of the total microseismic deformation with either number of events of total microseismic strength require proper normalization for detection levels and magnitude of completeness.

Table 2

Number of events and cumulative seismic moment

graphic
 
graphic
 
Fig. 6

Cumulative number of events versus magnitude for the two stages. The vertical green line is the magnitude of completeness, above which a linear power law relationship exists. For reference the blue line is the uniformly detected magnitude, and the vertical orange and red arrow indicates the smallest magnitude recorded in each stage from Fig. 5.

Fig. 6

Cumulative number of events versus magnitude for the two stages. The vertical green line is the magnitude of completeness, above which a linear power law relationship exists. For reference the blue line is the uniformly detected magnitude, and the vertical orange and red arrow indicates the smallest magnitude recorded in each stage from Fig. 5.

Finally, consider the design of a monitoring project to compare the fracture response of two or more different stimulation or completion designs. Prior to deciding the stages that are to be compared, it is critical to perform a pre-survey design study. Such design studies involve determining the minimum detectable magnitude and expected location accuracy at different positions (Maxwell et al. 2003a), along with potential assessment of the expected number of detected events (Maxwell, 2011b). The output of this design study is an expected range within which accurate microseismic results can be expected with a specific sensitivity. This range can then be used to identify the planned stages of the treatment well(s) that can be effectively monitored and compared. Fig. 7 shows two conceptual completion and stimulation designs that could be undertaken for such a multi-well project. The best scenario is to alternate designs within each well, and then apply the aforementioned comparative analysis. Not only are neighboring stages most favorable for comparison, but such a design allows comparison between wells which could otherwise be complicated with geologic variations under a scenario where different designs were executed in different wells. An unfavorable fracture comparison is also depicted where the stages in the toe and heel of the well are varied, in which case comparison of the more distant fracs would not be expected to result in high quality microseismic data. Comparative interpretation would then be compromised.

Fig. 7

Conceptual plot of an ideal treatment comparison (stages in blue box for top two wells) and non-ideal comparison (stages in red box for bottom two wells) for two different comparisons overlain with design limits from a pre-survey design study.

Fig. 7

Conceptual plot of an ideal treatment comparison (stages in blue box for top two wells) and non-ideal comparison (stages in red box for bottom two wells) for two different comparisons overlain with design limits from a pre-survey design study.

Microseismic Applications

Microseismic interpretations can be applied in both reservoir characterization and development. The reservoir development applications can be divided into three categories based on the timeframe of the application.

  1. Real-time Fracture Control

    • Diversion and re-fracturing

    • Identification of geo-hazards

    • Stage modification

  2. Completion Effectiveness

    • Completion efficiency

    • Fracture treatment design

    • Completion strategy

  3. Field Development

    • Well placement

    • Well spacing

    • Drainage patterns and recovery

Real-time applications of microseismic interpretations require fast and reliable event detection and location algorithms along with associated QC attributes and skilled engineers to interpret the images and make immediate recommendations for changes in the treatment execution. Stage modification might be better labeled "almost" real-time, as subsequent stages are modified based on the results of previous stages. Within a short time after the microseismic interpretation is complete, the interpretation can be used to evaluate the current completion strategy and modify future completions. However, fracture treatment design and field development applications of microseismic interpretations typically require a minimum of 6-12 months of production data. Fracture treatment design, staging and perforation strategy, and field development applications require integration with geological, geomechanical and geophysical data and more comprehensive engineering workflows.

The basic microseismic reservoir development application workflow is illustrated in Fig. 8, separating the applications that are based primarily on visualization of the microseismic interpretation and the applications that require modeling. Currently, Real-Time applications are based on visualization of the microseismic event patterns, ESV, event histograms, etc. Completion Effectiveness applications such as improving staging and perforating, and evaluating completion efficiency are also visualization-based applications. However, hydraulic fracture modeling and reservoir simulation are required for fracture treatment design and completion strategy applications (optimum number of perforation clusters and stages).

Fig. 8

Basic microseismic application workflow illustrating visualization based applications (brown) and modeling based applications (blue).

Fig. 8

Basic microseismic application workflow illustrating visualization based applications (brown) and modeling based applications (blue).

The remainder of the text provides more detail on both the reservoir characterization and development applications of microseismic interpretation, followed by a review of some of the most common misapplications of microseismic data to provide background and perspective to the application workflows.

Reservoir Characterization Using Microseismic Measurements

Integrating microseismic with geological earth models and seismic reservoir characterization to understand the impact of geology on the hydraulic fractures is a growing application. Changes in faulting, fractures, material properties and stress conditions can all lead to variations in the hydraulic fractures. Interpretation relies on integration of the microseismic interpretation with various other data including geologic models, seismic reservoir characterization, wireline borehole images, geomechanical earth models and hydraulic fracture injection data. The integrated microseismic interpretation can rely on either absolute or relative interpretation between stages and/or wells. Ultimately these studies could result in predictive capabilities to assess what fracture geometry will be created and identify ‘sweet spots’ for strategic well placement. The geological integration also aids understanding of the mechanism of anomalous microseismic patterns, including height growth, asymmetric or preferential growth in certain directions.

Fault activation can have a significant impact on the stimulation and can be investigated through microseismic attributes as well as geologic confirmation of the fault. Microseismic magnitude and b-values will change as a fault is activated during a hydraulic fracture. Furthermore, focal mechanisms can be used to distinguish microseismic deformation occurring along orientations consistent with known geologic faults. Geologic evidence of suspected faulting indicated by the microseismic interpretation can be detected using seismic ‘edge’ detection algorithms to find discontinuous reflectors in the seismic reflection data and help improve the resolution and detect subtle faults. Alternatively, geologic borehole evidence such as cores or wireline imaging can be used independently or supplementary to the seismic data. Obviously confidence in the microseismic fault activation interpretation is increased with independent geologic verification of the fault (Fig. 9). Ideally the fault would be identified beforehand, and the well placement and completion strategy adjusted to mitigate any risk of fault activation. Real-time geo-hazard control can be realized through populating the microseismic data as they occur into the geologic model, to validate the mitigation strategy.

Fig. 9

Left side shows results of an edge detection algorithm applied to a potential fault activation indentified through magnitude, b-values and mechanisms. Edge detection indicates a lineation associated with a distinct fault located below the reservoir (right side). In this example, the fault acts as a barrier and limits the hydraulic fracture growth to the NE (after Maxwell et al., 2011b).

Fig. 9

Left side shows results of an edge detection algorithm applied to a potential fault activation indentified through magnitude, b-values and mechanisms. Edge detection indicates a lineation associated with a distinct fault located below the reservoir (right side). In this example, the fault acts as a barrier and limits the hydraulic fracture growth to the NE (after Maxwell et al., 2011b).

Identification of fracture complexity is clearly topical, particularly in stimulation of shales. Fracture complexity (with fractures growing in multiple directions) requires pre-existing fractures in various directions and relatively small differences in the principal horizontal stresses. Interpretation of pre-existing fractures is similar to fault activation, although the utility of microseismic b-value and magnitude changes is not as clear. However, microseismic source mechanisms are useful to confirm deformation in directions consistent with the fractures. Additional seismic attributes such as anisotropy determined from either the microseismic or reflection data can further help detect the rock fabric. Evidence of isotropic stresses can further confirm geomechanical conditions favorable to generation of fracture complexity, as discussed next. Potential complexity could be identified prior to the stimulation treatment and the fracture design and completion strategy could then be implemented based on expected fracture geometries. Real-time monitoring can then be used to make refinements on a stage-by-stage basis. Careful filtering for the most accurate microseismic locations is a critical step, to avoid apparent complexity in the microseismic data associated with low confidence locations (Cipolla et al., 2011b).

The geomechanical assessment of stress anisotropy and heterogeneity uses microseismic workflows similar to fracture complexity. On the reservoir characterization side, however, additional integration of mechanical earth models based on logging results and hydraulic fracture pressure data can be used to quantify the stress state. Stress heterogeneity can be assessed through injection pressure, sonic logging and seismic reservoir characterization. For example, Daniels et al. [2007] and Rich and Ammerman [2010] describe assessment of stress anisotropy and heterogeneity for a Barnett frac where the microseismic data showed significant variability along the treatment well.Cipolla et al. [2010] describe a geomechanical reconstruction of the stress state that confirms the observed fracture geometry changes related to changes in the stress anisotropy (Fig. 10). In another case study, Maxwell et al. [2011a] describe identifying preferential fracture growth into regions of the reservoir identified as having low Poisson’s Ratio (PR) by amplitude-versus-offset analysis of seismic reflection data (Fig. 11). In this example, the variation in PR was interpreted to be related to changes in the stress state and confirmed by integrating with ISIP’s.

Fig. 10

Map view of microseismic recorded during a 4 stage treatment of a Barnett horizontal well (top) overlain on seismic curvature (contours) and seismic fabric direction (arrows) (after Rich and Ammerman, 2010). Bottom shows contours of stress anisotropy from a geomechanical reconstruction (after Cipolla et al., 2010). The first two stages are in a region of higher stress anisotropy and where the fabric is aligned with the maximum stress direction resulting in creation of a relatively planar fracture. The last two stages are in a region with more isotropic stresses, which along with the fabric more parallel to the well results in more fracture complexity.

Fig. 10

Map view of microseismic recorded during a 4 stage treatment of a Barnett horizontal well (top) overlain on seismic curvature (contours) and seismic fabric direction (arrows) (after Rich and Ammerman, 2010). Bottom shows contours of stress anisotropy from a geomechanical reconstruction (after Cipolla et al., 2010). The first two stages are in a region of higher stress anisotropy and where the fabric is aligned with the maximum stress direction resulting in creation of a relatively planar fracture. The last two stages are in a region with more isotropic stresses, which along with the fabric more parallel to the well results in more fracture complexity.

Fig. 11

Map view of PR contours overlain with microseismic data recorded during stimulation of three wells in the Montney Shale (after Maxwell et al., 2011). Yellow and red is low PR (~0.1) and blue and purple is high PR (~0.3). Also plots are ISIP gradients (MPa/m) for various stages. Note that the microseismic shows asymmetric fractures growing preferentially into the lower PR regions. ISIPs confirm these are regions with lower stresses.

Fig. 11

Map view of PR contours overlain with microseismic data recorded during stimulation of three wells in the Montney Shale (after Maxwell et al., 2011). Yellow and red is low PR (~0.1) and blue and purple is high PR (~0.3). Also plots are ISIP gradients (MPa/m) for various stages. Note that the microseismic shows asymmetric fractures growing preferentially into the lower PR regions. ISIPs confirm these are regions with lower stresses.

Real-Time Applications

The application of microseismic measurements in real- time is focused primarily on the location of the microseismic events, providing a measurement of the fracture top and/or bottom and the fracture location. The real-time applications are:

  • - Diversion or re-fracturing

  • - Identifying and/or avoiding geo-hazards

  • - Height growth control

  • - Stage modification

The basic workflow for real-time microseismic applications is shown in Fig. 12. The data requirements for real-time microseismic applications are usually much less than for other applications, but a basic geologic model is normally required for most real-time applications. The input parameters required for real-time interpretation consist of event locations, number of events, magnitude, stimulated volume, b-value, and event histograms. The applications are primarily based on visual observations of microseismic behavior to identify the location of the hydraulic fracture. Based on the real-time visualization of microseismic behavior, the treatment can be modified to achieve the desired geometry or avoid geo-hazards.

Fig. 12

Basic workflow for real-time microseismic mapping

Fig. 12

Basic workflow for real-time microseismic mapping

Diversion and Re-fracturing

It is often desirable to control where the fracture propagates, but very difficult in practice to successfully apply diversion technology to effectively control fracture growth. Pressure measurements alone are typically not sufficient to reliably evaluate diversion effectiveness and fracture location. Thus microseismic mapping has become the primary tool to evaluate real-time treatment control. It is important to note that real-time applications such as diversion and re-fracturing rely only on the relative position of the microseismic events at various stages in the treatments and, as described in the comparative interpretation section, are less sensitive to uncertainty in the event locations.

Fig. 13 shows how real-time MSM can be used to control the location of the hydraulic fracture using fiber diversion technology to improve stimulation effectiveness (Daniels et. al. 2007). The initial stimulation on this well was monitored using MSM, but real-time control technologies were not yet being routinely implemented and the fracture network coverage of the lateral was inefficient, with about 30% of the lateral un-stimulated (Fig. 13). The figure shows the interpretation tools used for real-time applications; histograms of event locations to locate the primary depth of fracture propagation in the lateral (upper right corner) and stimulated volume (ESV in this case, upper left corner) for each re-fracture stage to identify the volumetric location of fracture growth. Fig. 13 shows the microseismic event distribution for the initial fracture treatments and the re- fracture treatment, illustrating the un-stimulated and under- stimulated regions after the initial fracture treatment. Stage 1 of the re-fracture treatment did not contain any diverting material and propagated in the same region as the initial fracture treatment (Fig. 13, upper right graph). Stage 2 contained diverting material in a modest concentration and failed to change the location of fracture propagation (Fig. 13, bottom left graph). Based on the real-time application of the microseismic interpretation, more aggressive concentrations of diverting material were pumped in stage 3, resulting in significant improvement in the stimulation coverage (Fig. 13, lower right graph). The production improvement after the re- stimulation is shown in Fig. 14. Initial well productivity more than doubled and longer term production improved by 60%.

Fig. 13

Case history #3, Real-Time diversion using MSM. Initial fracture treatment was did not effectively stimulate the lateral (upper left graph). Re-fracture treatment stage 1 (yellow dots) propagated in the same portion of the lateral as the initial treatment (upper right graph); as did the first diversion attempt, stage 2 (blue dots, bottom left graph). More aggressive diversion allowed stage 3 (red dots) to direct fracture propagation to un-stimulation portion of the lateral and more effectively stimulate the under- stimulated reservoir regions (lower right graph). Histogram charts show the location of events along the lateral for each re- fracture stage (upper right), while estimated stimulated volume or ESV for each re-fracture stage is shown in the upper left of each graph.

Fig. 13

Case history #3, Real-Time diversion using MSM. Initial fracture treatment was did not effectively stimulate the lateral (upper left graph). Re-fracture treatment stage 1 (yellow dots) propagated in the same portion of the lateral as the initial treatment (upper right graph); as did the first diversion attempt, stage 2 (blue dots, bottom left graph). More aggressive diversion allowed stage 3 (red dots) to direct fracture propagation to un-stimulation portion of the lateral and more effectively stimulate the under- stimulated reservoir regions (lower right graph). Histogram charts show the location of events along the lateral for each re- fracture stage (upper right), while estimated stimulated volume or ESV for each re-fracture stage is shown in the upper left of each graph.

Fig. 14

Production date for case history #3 showing a 60% production improvement after re-fracture treatment.

Fig. 14

Production date for case history #3 showing a 60% production improvement after re-fracture treatment.

Identifying and Avoiding Geo-Hazards, Height Growth Control

Fracturing into geo-hazards such as water-productive zones, faults, and karsts can significantly reduce stimulation effectiveness. The real-time application of microseismic can be used to identify and avoid geo-hazards. When identifying geo-hazards, the event magnitude and the temporal magnitude distribution are used in conjunction with event locations.Fig.15 is an example of downward fracture height growth into a water-productive zone due to fault activation. The visualization of the spatial distribution of microseismic events clearly shows an anomalous pattern. Maxwell et al. [2008], Warpinski [2009], and Maxwell et al. [2011a] show that fault activation during a hydraulic fracture treatment can be identified by the magnitude and location of microseismic events. Fig. 16, from Maxwell et al. [2009], shows how event magnitudes can be used to identify fault activation. Combining the event magnitude and the event locations provides a reliable real-time tool to identify fault activation. Maxwell at al. [2009] and Downie et al. [2010] show how the distribution of event magnitudes or b-values can provide an additional diagnostic to identify fault activation. Fig. 17, from Maxwell et al. [2009], illustrates how real-time application of b-value calculations can be used as an indicator of potential fault activation. Microseismic event distributions during a typical hydraulic fracture treatment exhibit a b-value of approximately two, while the b-value during episodes dominated by fault activation is around one. Real-time calculations of b-values using a moving average time window may provide a tool to confirm fault activation much earlier in the treatment. In this illustration, the deviation of the b-value from two around 125 minutes would be a warning sign of fault activation.

Fig. 15

Example of fault activation. Microseismic mapping shows a high concentration of large magnitude events with a spatial distribution indicating the hydraulic fracture contacted a fault.

Fig. 15

Example of fault activation. Microseismic mapping shows a high concentration of large magnitude events with a spatial distribution indicating the hydraulic fracture contacted a fault.

Fig. 16

Example of fault activation identified using event magnitude (from Maxwell et al., 2009).

Fig. 16

Example of fault activation identified using event magnitude (from Maxwell et al., 2009).

Fig. 17

llustration of real-time b-value calculations to identify fault activation (from Maxwell et al., 2009).

Fig. 17

llustration of real-time b-value calculations to identify fault activation (from Maxwell et al., 2009).

In many cases geo-hazards are already identified based on petrophysical or seismic measurements and the application of real-time microseismic measurements simply requires the integration of the geologic model with the microseismic locations. Most real-time visualization software can provide this integration. The real-time visualization of event locations and magnitudes (events symbols are sized proportional to the event magnitude) can then be used to recognize that the hydraulic fracture is propagating into a geo-hazard. Once propagation into a geo-hazard is identified there are a number of options to modify the treatment including aborting the treatment, reducing the injection rate, and the application of diversion technologies (fiber, proppant slugs, viscous pills, etc.). Waters et al. [2009b] provide an excellent example of using real-time microseismic interpretations to adjust injection rate and apply diversion technology to avoid a fault. King et al. [2008] discuss techniques to control downward growth into a water-productive zone that incorporated real-time microseismic measurements. In many cases the most prudent option is to abort the current stimulation treatment and proceed to the next stage to minimize the risk of water production or economic waste (i.e. – fracturing into a fault or karsts). Unless the option to use diversion technology was planned in advance of the treatment, the real-time application of microseismic mapping is usually limited to adjusting the injection rate or aborting the treatment.

Completion Effectiveness Applications

Stage modification might be considered a "near" real-time application of microseismic mapping, as the treatment design and perforating strategy are changed in subsequent stages based on the microseismic interpretation of previous stages. The primary interpretations required are fracture location and geometry and stimulated volume. Based on the microseismic interpretations, the perforating strategy and/or treatment design may be changed on subsequent stages to improve vertical or lateral wellbore stimulation and completion efficiency. Ejofodomi et al. [2010] illustrate how near real-time microseismic mapping can be used to modify staging and perforating in vertical wells in West Texas, while Fisher et al. [2004] illustrate this application in horizontal Barnett shale wells. The details of real-time stage modification are a subset of applications that target improving completion effectiveness (discussed next).

Completion Effectiveness Applications

Microseismic mapping can be applied to improve completion effectiveness, resulting in more efficient and productive completions. The application of microseismic interpretations to improve completion effectiveness consists of both visualization-based applications and model-based applications (Fig. 8).

Completion effectiveness is controlled by three factors, the completion strategy, the stimulation design, and the completion efficiency, with completion efficiency and completion strategy closely linked. Most unconventional resources are developed using multiple fracture treatments with multiple perforation "clusters" in each fracture treatment stage. The stage and cluster geometry is "designed" to evenly distribute stimulation fluid and proppant to each cluster, in an attempt to create separate hydraulic fractures at each perforation cluster in each stage. There is a tradeoff between the number of stages, number of perforation clusters per stage, spacing between perforation clusters, well productivity, and completion cost.

Completion efficiency is controlled by the connection between the reservoir and the wellbore and the stimulation coverage of the target interval for a given completion strategy and stimulation design. The completion strategy consists of parameters such as the distance between stages, number of perforation clusters, and/or distance between perforation clusters. Even a good completion strategy can be ineffective if the completion efficiency is low due to such issues as poor cement.

For horizontal well completions, increasing the number of fracture treatment stages for a given lateral length will increase productivity, but completion costs will also increase. There is typically a compromise between the spacing between clusters, number of stages, and cost. For a given spacing and treatment volume per perforation cluster, increasing the number of perforation clusters per stage will reduce the number of stages required, reducing completion time and cost. However, production logs show that 25% or more of the perforation clusters in horizontal shale-gas completions are not producing (Miller et al., 2011). In addition, this data shows that the percentage of non-productive perforation clusters increases as the number of perforation clusters is increased.

Completion Efficiency

The visualization-based applications are primarily focused on improving the completion efficiency, which include evaluating stage isolation techniques such as cemented versus un-cemented completions, identifying cement quality issues and assessing perforation cluster spacing and location and stage overlap. The workflow for these visualization-based applications is essentially the same as that for real- time applications (Fig. 12), except that the final step is now "Completion Modification" instead of "Treatment Modification". Fig. 18 illustrates how stimulated volume calculations can be visualized to evaluate completion efficiency in a horizontal shale- gas completion (Daniels et al., 2007). In this example, the four stage stimulation strategy did not stimulate the entire lateral, as evidenced by the portions of the lateral that do not exhibit microseismic activity. In most shale-gas reservoirs, the drainage area is limited to the proximity of the hydraulic fracture, thus large areas with no microseismic activity can be assumed to be non-productive. Increasing the number of stages on subsequent wells will improve completion efficiency. Fig. 19 is an example of applying microseismic interpretations to evaluate completion efficiency in a tight gas reservoir (Baihly et al., 2009). This example shows stage overlap due to poor cementing in a plug- and-perf completion and multiple fractures propagating from isolation packers in an un- cemented "ball-drop" completion.

Fig. 18

Example of stimulated volume showing less than optimum lateral coverage and opportunities to improve completion efficiency. (after Daniels et al., 2007).

Fig. 18

Example of stimulated volume showing less than optimum lateral coverage and opportunities to improve completion efficiency. (after Daniels et al., 2007).

Fig. 19

Example of the application of microseismic mapping to evaluate completion efficiency in a tight gas well (after Baihly et al., 2009). Left graph shows stage overlap in a case and cemented completion (plug & perf). Right graph shows two fractures initiating for near isolation packers in an un-cemented completion (external casing packers with frac ports actuated using balls).

Fig. 19

Example of the application of microseismic mapping to evaluate completion efficiency in a tight gas well (after Baihly et al., 2009). Left graph shows stage overlap in a case and cemented completion (plug & perf). Right graph shows two fractures initiating for near isolation packers in an un-cemented completion (external casing packers with frac ports actuated using balls).

Fig. 20 shows an example of stage overlap in a vertical tight gas well. In this example stages 1, 2, and 3 stimulate essential the same vertical section. Subsequent production logging measurements indicated that very little gas was being produced from stages 2 and 3. In this example, eliminating stages 2 and 3 and pumping a larger treatment with perforations in the lower portion of the interval would reduce completion costs and provide similar production.

Fig. 20

Illustration of stage overlap in a vertical well (after Peterman et al., 2005). In this example, the bottom three stages overlap, suggesting they can be combined into a single stage with a larger fracture treatment covering the entire interval. A production log shows that most of the gas in coming from stage 1, indicating the location of the perforation clusters should be in the lower portion if the interval. Fracture half-length, fracture conductivity, and permeability are shown in the table for reference.

Fig. 20

Illustration of stage overlap in a vertical well (after Peterman et al., 2005). In this example, the bottom three stages overlap, suggesting they can be combined into a single stage with a larger fracture treatment covering the entire interval. A production log shows that most of the gas in coming from stage 1, indicating the location of the perforation clusters should be in the lower portion if the interval. Fracture half-length, fracture conductivity, and permeability are shown in the table for reference.

These examples illustrate how simple visualization-based MS applications can be used to evaluate completion efficiency and identify areas where future completion strategies and operational procedures can be improved.

Completion Strategy and Stimulation Design

Microseismic interpretations can be used to improve completion strategies and stimulation designs. These applications are much more modeling intensive and can require additional data and measurements. The workflow for completion strategy and fracture design applications of microseismic mapping is shown in Fig. 21. This workflow requires fit-for-purpose hydraulic fracture and reservoir simulation models and the ability to efficiently integrate multi-discipline data and interpretations. The workflows used when improving completion strategies and stimulation designs are very similar to those used for field development applications of microseismic interpretations (discussed later). The primary difference is that the field development applications are focused on calibrating the reservoir simulation model and determining drainage architecture, while completion effectiveness applications are focused on calibrating the hydraulic fracture model and improving the completion strategy. Therefore, the integration of production data and production history matching are not included in the completion effectiveness workflow.

Fig. 21

Workflow for completion strategy and stimulation design applications of microseismic mapping.

Fig. 21

Workflow for completion strategy and stimulation design applications of microseismic mapping.

Applying microseismic interpretations to improve stimulation designs and completion strategies starts with a detailed Earth Model, which includes integration of both seismic and wellbore measurements. As discussed, microseismic measurements can also be an important input into seismic workflows, providing insights into the relationships between hydraulic fracture propagation, natural fractures and faults, structure, and rock properties. The primary application of microseismic interpretations in this workflow is for fracture model calibration (reference Fig. 21). The appropriate hydraulic fracture model will depend on the level of fracture complexity, with planar fracture models being appropriate when fracture growth is relatively simple. However, complex fracture models are required for many shale completions. The evaluation of fracture complexity and the calibration of complex fracture models are discussed by Cipolla et al. [2011a]. Determining the degree of hydraulic fracture complexity is an important application of microseismic measurements and a critical component of stimulation design (Cipolla et al., 2008a, 2008b).

The workflow shown in Fig. 21 assumes that the completion strategy will be determined independently, using simple geometric spacing of perforation clusters and stages or more sophisticated algorithms that account for variations in reservoir quality, rock properties and stress, location of nature fractures and faults, and wellbore configuration (Cipolla et al., 2011c). Production logs, if available, can provide important measurements to evaluate completion efficiency and calibrate completion strategy algorithms. Comparing production log results to fracture modeling predictions of near-wellbore conductivity may improve the fracture model calibration.

The primary microseismic interpretations required for fracture model calibration and improving completion strategy are fracture location and either geometry (e.g. height and length) or stimulated volume, depending on the degree of fracture complexity.

Example Application. A horizontal Barnett shale completion is used to illustrate the application of microseismic measurements for fracture model calibration and stimulation design. This example draws from the work of Daniels et al. [2007], Rich and Ammerman [2010] and Cipolla et al. [2010], This well has a rich dataset, including microseismic monitoring for all stages of the stimulation, advanced sonic logs that provided estimates of minimum and maximum horizontal stress and 3D seismic interpretations of curvature and natural fracture orientations (Rich and Ammerman 2010). The 3200-ft lateral was drilled in the direction of minimum horizontal stress (σh), encouraging transverse hydraulic fractures, and four slickwater fracture treatments were pumped. Each treatment consisted of 25,000 bbls of water and 440,000 lbs of proppant. The fracture treatments were monitored using an array of geophones in an offsetting wellbore. In this example the microseismic image has been patterned after the results presented by Daniels et al. [2007], with the example focused on two of the four stages for brevity. The microseismic images patterned after the stage 1 and 3 fracture treatments are shown in Fig. 22, illustrating relatively linear event patterns in stage 1 and a mixture of complex and linear event patterns in stage 3. The difference in event patterns was explained by variations in stress regime and natural fracture development (Daniels et al., 2007;,Rich and Ammerman, 2010;,Cipolla et al., 2010).

Fig. 22

Microseismic event patterns for two fracture treatment stages in a Barnett Shale completion. Based on the results presented by Daniels et al. [2007].

Fig. 22

Microseismic event patterns for two fracture treatment stages in a Barnett Shale completion. Based on the results presented by Daniels et al. [2007].

Fig. 23

Discrete Fracture Network (DFN) representing the distribution of natural fractures in the vicinity of the example horizontal well. Note that the dominant natural fractures trending NE-SW in the toe section of the lateral are absent in the heel section.

Fig. 23

Discrete Fracture Network (DFN) representing the distribution of natural fractures in the vicinity of the example horizontal well. Note that the dominant natural fractures trending NE-SW in the toe section of the lateral are absent in the heel section.

The combination of the 3D seismic interpretation, advanced sonic log, and microseismic measurements provide a much more reliable understanding of differences in hydraulic fracture growth along the lateral and constrain the DFN and Earth Model. A detailed geologic description of the natural fractures was not available and the DFN was generated using a stochastic model guided by the advanced seismic interpretation and microseismic measurements (Fig. 23). Minimum horizontal stress in the Barnett layer varied from 0.70 psi/ft in the toe-section of the lateral to 0.62 psi/ft in the heel section, while maximum horizontal stress ranged from 0.77 psi/ft to 0.65 psi/ft in the toe and heel, respectively. The difference in the two horizontal stresses varied from 0.10 psi/ft in the toe section of the lateral to 0.03 psi/ft in the heel.

The complex hydraulic fracture propagation model (Weng et al., 2011) was calibrated to the microseismic image (Fig. 22) for the two stages. The calibration consisted of minor adjustments in the maximum horizontal stress and the evaluation of numerous realizations of the DFN. The complex fracture geometry is compared to the microseismic image in Fig. 24, showing good agreement. As noted, the fracture model calibration assumes that the hydraulic fracture geometry is defined by the microseismic volume. Although not shown, the fracture was mostly contained in the reservoir interval. Fig. 25 shows a three dimensional view of the stage 1 complex fracture geometry, illustrating that a large portion of the fracture network is not propped. A large portion of the stage 3 fracture geometry was also not propped.

Fig. 24

Comparison of complex hydraulic fracture geometry and microseismic image (plan view).

Fig. 24

Comparison of complex hydraulic fracture geometry and microseismic image (plan view).

Fig. 25

Stage 1 fracture geometry showing proppant and un- propped regions of the fracture network.

Fig. 25

Stage 1 fracture geometry showing proppant and un- propped regions of the fracture network.

The complex fracture modeling results indicated that un-propped fracture conductivity will play a critical role in well productivity. It is often assumed that propped fracture conductivity in shale plays is essentially infinite with respect to the very tight matrix permeability; however the relationships between propped fracture conductivity, un- propped fracture conductivity, fracture complexity, and well productivity are complex and not adequately described using classical fracture conductivity theory developed for planar hydraulic fractures (Cipolla et al., 2008).

Evaluating the impact of variations in fracture treatment designs, propped and un-propped fracture conductivity, and well productivity requires detailed reservoir simulation. Although reservoir simulation is included in the completion effectiveness workflow, this portion of the workflow will be discussed in the following section for expediency. However, even without reservoir simulation and production forecasting, the complex fracture modeling results indicate that treatment design changes that increase propped fracture area are warranted.

Once the fracture model is calibrated and the stimulation design has been "optimized", the final step in the fracture design and staging strategy workflow is to evaluate the impact of stage spacing and number of perforation clusters per stage on well productivity and completion cost. As with fracture design, this step requires an efficient integration of fracture modeling and reservoir simulation results and will be discussed in the context of the field development workflow.

Field Development Applications

The field development applications of microseismic mapping are focused on evaluating well productivity to determine drainage architecture and hydrocarbon recovery. The basic workflow for field development applications (Fig. 26) is very similar to the completion effectiveness workflow, but focuses on calibrating the reservoir simulation model by history matching well performance. However, fracture model calibration remains in the field development workflow, as reservoir simulation history matching results may indicate changes in proppant distribution, requiring additional fracture model calibration.

Fig. 26

Workflow for field development applications of microseismic mapping.

Fig. 26

Workflow for field development applications of microseismic mapping.

The example presented in the previous section will be evaluated in further detail using reservoir simulation to determine the effective fracture area or effective stimulated volume. Prior to the introduction of complex hydraulic fracture models, it was difficult to estimate how much of the stimulated volume calculated from the microseismic image was actually productive (Fig. 27). In addition, there was no direct "link" between the fracture treatment design, stimulated volume, and production evaluation. With the introduction of complex hydraulic fracture models, there is now an opportunity to perform much more rigorous evaluation. However, two obstacles must be overcome; (1) efficient integration of microseismic data, complex fracture models, and reservoir simulation models along with geological, petrophysical, and geophysical interpretations on a common software platform and (2) automated gridding routines to generate reservoir simulation grids that exactly represent the fracture geometry and distribution of fracture conductivity predicted by the complex fracture model. These two obstacles have been addressed with the recent introduction of seismic-to-simulation software for unconventional reservoirs (Cipolla et al., 2011d).

Fig. 27

Stimulated volume calculated from the microseismic image (shaded area). But how much is actually productive?

Fig. 27

Stimulated volume calculated from the microseismic image (shaded area). But how much is actually productive?

The complex hydraulic fracture geometry for stages 1 and 3 (Fig. 24) was automatically gridded and input into the reservoir simulation model. Previous history matching work using a much less sophisticated complex fracture model was presented by Cipolla et al. [2010] and indicated a propped fracture conductivity of 15 md-ft and an un-propped fracture conductivity of 0.03 md-ft. As a reference, un-propped and partially propped fracture conductivity values are presented in Fig. 28, with the range of closure stress for this example highlighted. The history match value of 0.03 md-ft falls on the low end of the range. These values will be used as a starting point to illustrate the impact of un- propped conductivity on productivity, drainage area, and gas recovery. Table 3 shows the basic reservoir properties used in the reservoir simulations.

Fig. 28

Un-propped and partially propped fracture conductivity (adapted from Fredd et al., 2001).

Fig. 28

Un-propped and partially propped fracture conductivity (adapted from Fredd et al., 2001).

Table 3

Reservoir properties

top depth 6900 ft 
reservoir thickness 400 ft 
reservoir pressure 3200 psi 
porosity 3% 
permeability 100 nd 
water saturation 30 % 
gas gravity 0.6 
top depth 6900 ft 
reservoir thickness 400 ft 
reservoir pressure 3200 psi 
porosity 3% 
permeability 100 nd 
water saturation 30 % 
gas gravity 0.6 

The pressure distribution after 20 years of production is shown in Fig. 29. The simulations were performed using a constant flowing bottomhole pressure of 1000 psiThe drainage architecture is dominated by the propped regions of the fracture network (blue regions), with limited drainage outside the propped regions (yellow/orange regions). The projected gas recovery is 3.5 BCF in 20 years. There is very little pressure depletion at the un-propped extremities of the fracture network, even after 20 years. Fig. 30 compares the microseismic stimulated volume (Fig. 27) and the 20-year pressure depletion, illustrating that the effective stimulated volume is much less that the microseismic stimulated volume due to the un-propped portions of the fracture network. This illustrates the importance of integrating complex fracture modeling and reservoir simulation, as the distribution of propped and un-propped conductivity will dictate productivity and drainage patterns. In this case, un-propped conductivity appears very low, limiting drainage in the un-propped regions of the network.

Fig. 29

Pressure distribution after 20-yrs. Propped FC = 15 md-ft, Un-propped FC = 0.03 md-ft. Cumulative gas production

Fig. 29

Pressure distribution after 20-yrs. Propped FC = 15 md-ft, Un-propped FC = 0.03 md-ft. Cumulative gas production

Fig. 30

Comparison of microseismic stimulated volume and 20 year pressure depletion. The effective stimulated volume is much less that the microseismic stimulated volume due to un-propped portions of the fracture network.

Fig. 30

Comparison of microseismic stimulated volume and 20 year pressure depletion. The effective stimulated volume is much less that the microseismic stimulated volume due to un-propped portions of the fracture network.

The impact of un-propped fracture conductivity on well productivity and gas recovery for un-propped conductivities ranging from 0.0003 md-ft to 0.3 md-ft is shown in Fig. 31. There is a dramatic difference in the effective stimulated area and gas drainage depending on the un-propped conductivity. Simulations were also performed to evaluate the impact of increasing propped fracture conductivity and the results indicated that 15 md- ft was essentially infinite conductivity. Fig. 32 compares the cumulative gas production for un-propped fracture conductivities of 0.0003, 0.03, and 1 md-ft, showing that even an un-propped conductivity of 0.03 md-ft contributes significantly to well productivity and gas recovery. This is illustrated by comparing the gas recovery when un-propped conductivity is 0.0003 md-ft to that of the history match value of 0.03 md-ft; without the contribution of the un-propped regions of the fracture network there is a 40% decrease in gas production. Conversely, modifying fracture treatment designs to prop more of the fracture network could result in significant improvements in well productivity and gas recovery, as evidenced by the 90% increase in gas recovery when fracture conductivity is increased to 1 md-ft in the un-propped regions of the fracture network. When un-propped fracture conductivity is "relatively" high, the effective stimulated volume approaches the microseismic volume.

Fig. 31

Comparison of 20-year pressure depletion for un-propped conduct ivies of 1 md-ft (left) and 0.0003 md-ft (right).

Fig. 31

Comparison of 20-year pressure depletion for un-propped conduct ivies of 1 md-ft (left) and 0.0003 md-ft (right).

Fig. 32

Comparison of 20-year cumulative production for un-propped fracture conductivities of 0.0003 md-ft, 0.03 md-ft and 1 md-ft.

Fig. 32

Comparison of 20-year cumulative production for un-propped fracture conductivities of 0.0003 md-ft, 0.03 md-ft and 1 md-ft.

The un-propped fracture conductivity presented in Fig. 28 is based on laboratory experiments using relatively high modulus core. Un-propped fracture conductivity can be significantly less for lower modulus rock (Cipolla et al., 2008). In addition, exposure to, and ineffective cleanup of, stimulation fluids may further reduce un-propped fracture conductivity. Therefore the importance of un-propped fracture conductivity on well productivity and drainage area could differ dramatically depending on the geological and geomechanical environment.

This example illustrates the application of microseismic mapping to determine drainage architecture and hydrocarbon recovery. The results show that the effective stimulated volume can be significantly less than the microseismic volume. In addition, un-propped fracture area can contribute significantly to well productivity and hydrocarbon recovery. In this example, well spacing is dictated primarily by the propped regions of the fracture network. Increasing the percentage of the fracture that is propped or partially propped could significantly improve well productivity and increase drainage area.

Well Spacing and Placement

The results of the previous example are used to illustrate the application of microseismic mapping for well spacing and placement. This is probably one of the most important issues that impacts recovery factor and economics for unconventional reservoir development. A detailed well spacing and placement evaluation is beyond the scope of this text, but the application will be illustrated using drainage patterns based on the previous reservoir simulations.

Fig. 33

Applying microseismic mapping for well placement and spacing

Fig. 33

Applying microseismic mapping for well placement and spacing

Misapplication of Microseismic Interpretations

It is widely assumed that the observed microseismic events directly represent the hydraulic fracture geometry. However, most of the microseismic events are only indirectly associated with the actual hydraulic fracture and are thus a proxy for the hydraulic fracture. This section highlights the implications of this proxy effect and provides some guidelines to convert the microseismic map into an accurate representation of the hydraulic fracture behavior. Hydraulic fracture operations typically involve the pumping of large volumes of fluid (typically 100,000 gal to 1,000,000 gal or more), while the total microseismic event volume (calculated from the cumulated magnitude of the events) is less than 1 barrel, as shown in Table 1. Clearly, the pumped volume cannot be stored only in the volume created by microseismic events. In addition, the energy released by the microseismic events is only a tiny fraction (e.g. - 1 ten thousandth) of the energy delivered by the pumps (Maxwell et al., 2008). This non-seismic energy is consumed as friction losses and tensile opening of the hydraulic fracture planes. Finally, the majority of observed microseismic events represent shear slip, which acts as a proxy for the tensile hydraulic fracture opening (Maxwell and Cipolla, 2011). Sources of shear movement include shear stresses in the fracture process zone around the tip, dog-legs in the generally planar fracture surface, and lubrication of pre-existing discontinuities by fluid leakoff.

Table 1 shows the calculated volume, area and energy of the microseismic events for the case study presented previously in the text. The microseismic volume and energy are a tiny fraction of the total volume and energy input of the hydraulic fracture treatment, and the area is a slightly larger fraction (but still less than 1%) of the calculated hydraulic fracture surface area. This highlights the concept of the microseismic events as a proxy for the actual hydraulic fracture network. Some common misapplications of microseismic images are now highlighted.

Moment Tensor Inversion: As discussed previously, Moment Tensor Inversion (MTI) can be used to infer the mechanism of the microseismic events. In some monitoring configurations (e.g. - single downhole monitor well), this inversion is non- unique. Even when it is unique (e.g. - from multiple downhole, surface or near-surface monitoring configurations) it is critical to understand that the source mechanism (opening, closing or shear) represents the deformation associated with the event, not the hydraulic fracture. For example Prince et al. [2011] suggest that if the cumulative volume strain from events during a treatment is negative, this net closing will result in reduced productivity in the reservoir. However, this statement would only be valid if the events in fact represented the hydraulic fracture network after both treatment and flowback. Furthermore, the example indicates predominant fracture closure while fluid is still being injected; suggesting that if the MTI is correct, aseismic fracture opening must be the dominant factor for mass conservation. Any deformation during pumping could very well be reversed during flowback and subsequent production. Perhaps more importantly, the only conductivity of interest relies on the residual deformation (e.g. - Moos et al., 2011). It should also be noted that shear deformation of real natural fractures results in some opening and residual conductivity (Fredd et al., 2001). Maxwell and Cipolla [2011c] show schematically how the shear deformation could be associated directly with the HF, and also how subsequent events could be associated with repeated deformation on the same area (Maxwell and Cipolla, 2011c, Fig. 10 and 11). For a very simple case of a pressurized crack, Nagel and Sanchez (2011) showed how shear strain (and hence microseismic events) can often be associated with deformation away from the fracture itself. It is thus essential to use a fracture model which accounts for the interaction of hydraulic fractures and natural fractures to simulate the development of the fracture network during stimulation (e.g.-Weng et al., 2011), and to calibrate this model with the microseismic interpretation (e.g.- Cipolla et al., 2011a), rather than to assume the microseismic interpretation itself represents the fracture network.

!Stimulated Volume: The concept of Stimulated Volume (SV) has evolved from simply drawing a rectilinear volume around the events (Mayerhofer et al, 2008), through approaches which exclude small or isolated events (Daniels et al., 2007), to algorithms which attempt to weight different parts of the stimulated volume differently, depending on the cumulative magnitude of the microseismic events. The last approach is based on work originally published by Maxwell et al., (2003b), and assumes that seismic deformation enhances permeability, which in turn leads to improved production. It is important to recognize that while SV is a valid qualitative measure of disturbed rock volume, it is a purely visual concept. These more advanced calculations of SV implicitly assume that more microseismic events imply better production. While this assumption may be valid in very specific circumstances, it is not generally true. It should be obvious from Figs Fig. 27 to Fig. 31 that production is very dependent on the fracture surface area, connectivity and conductivity. An individual microseismic event can only contribute to production if it results in area, connectivity and conductivity, none of which are captured in the calculation of SV. Furthermore, it is easy to overestimate the stimulated volume by ignoring event location uncertainty. The tempting assumption that "more events are better" encourages the use of lower magnitude events and/or lower SNR events. Typically these have higher location uncertainty, further increasing the volume over-estimation. The workflow to obtain a more realistic SV is outlined by Cipolla et al (2011b).

Underestimating hydraulic fracture dimensions: Assuming the SV is accurately determined by appropriately filtering of the event set and accounting for location error, it is commonly assumed that the fracture network lies entirely within the SV. However, it is important to account for distance bias. It is possible that some parts of the stimulation occurred too far from the observation well to be detected. Cipolla et al (2011b) explain how the magnitude-distance plot can be used to determine whether all events have been observed, and also how to eliminate the effects of distance bias. Unless there is a good explanation for the observed lack of symmetry, such as geological variations or interaction with a depleted zone, or the magnitude-distance plot clearly shows that all events would have been observed, any asymmetry should be assumed to be an artifact. The geophysical QC process also enables the identification of nodal planes which could explain unobserved events.

Furthermore, since the majority of the fracture tensile opening is slow (and hence aseismic), it cannot be assumed that lack of microseismic events implies there is no fracture. Although the events do not represent the actual fracture, they generally outline the location of the fracture, as discussed previously. There are some exceptions, such as activation of a fault, but these exceptions can generally be identified via other characteristics of the microseismic events, such as the b-value and changes in magnitude and source orientation (discussed previously).

Misinterpreting fracture complexity:As shown by Cipolla et al (2011b), event location uncertainty may result in apparent complexity, i.e. the cloud of events being assumed to represent complexity when in fact the events could equally well lie in a plane. Even if the events do not lie exactly in a plane, they could be the result of slip on natural features near the hydraulic fracture. There are several obvious mechanisms to trigger slip close to, but not exactly on, the fracture plane, such as leakoff, which may lubricate a natural feature, reducing the friction coefficient, or stress changes causing additional shear stress on these nearby features.

Ignoring mass balance and fracture mechanics:As noted previously, the microseismic volume is a small fraction of the total pumped volume. As such, any representation of the hydraulic fracture network must account for aseismic tensile opening and storage of the massive volumes of fluid and proppant pumped. In addition, fractures propagate perpendicular to the minimum stress, and are affected by natural fractures and other rock fabric. Hence the simulation of hydraulic fracture propagation, accounting for geology, rock properties, and stress, is essential to explain any microseismic interpretation.

Interpretation of overlapping microseismic volumes between different stages or wells:If the concept of SV is used without consideration of fracture conductivity, it is possible to assume that overlapping microseismic volumes imply inefficient stimulation. For example, it may be stated that some region is over-stimulated or that no more SV is being created. However, if the fracture networks are not fully conductive, such as the extremities of those in Fig. 31, then overlap is necessary to generate conductivity and access the resource.

The misapplications discussed in this section highlight three critical areas that must be considered for the proper application of microseismic interpretations:

  1. Geomechanics, specifically the increase in shear stress on natural fractures during and after stimulation,

  2. Complex Hydraulic Fracture simulation, accounting for the creation of connected surface area, proppant distribution, and other conductivity, and

  3. Explicit representation of hydraulic fracture planes in production simulation.

Summary

The keys to applying microseismic measurements are high quality geophysical processing, reliable and consistent interpretations and the integration of G&G data, fit-for-purpose fracture models, and reservoir simulation. In addition to engineering applications, microseismic measurements have reservoir characterization applications, providing important insights into natural fracture characteristics and stress regime. The complex workflows for Completion Effectiveness and Field Development applications of microseismic interpretations are now practical with the recent introduction of complex hydraulic fracture models, automated reservoir simulation gridding routines, and common software platforms where diverse datasets, interpretations, and models can be shared by engineers and geo-scientists.

Advanced processing of the microseismic waveforms holds great promise to provide significant insights into natural fracture characterization and to constrain hydraulic fracture models. However, microseismic deformation is a very small fraction of the total deformation and current interpretations cannot provide reliable insights into hydraulic fracture area, which is one of the key components that dictate well productivity. Nevertheless, changes in the microseismic response during hydraulic fracturing can be used to indentify fault activation, and help constrain the relative deformation predicted by geomechanical models. Attempting to define hydraulic fracture "opening" or "closing" using advanced processing such as moment tensor inversion (MTI) can be fraught with error and lead to misapplications. Therefore, current applications of advanced microseismic processing are limited primarily to reservoir and geomechanical characterizations (e.g. – natural fractures, faults, stress regime).

A number of important issues associated with advanced processing and stimulated volume interpretations impact the application of the results. Microseismic mapping may help define the location/distribution/orientation of the natural fractures, but cannot be used to characterize hydraulic fracture area or volume (not to be confused with MS stimulated volume). There is a significant potential for misapplication of advanced processing and SV interpretations if application workflows omit the following components or fail to recognize the limitations of the results.

  • Mass balance and fracture mechanics are required to estimate the fracture geometry.

  • Fracture models are required to estimate the location and concentration of proppant.

  • Geomechanics (e.g. – frac conductivity versus closure stress) is required to estimate the conductivity of the propped and un-propped regions.

  • Discretely modeling of the hydraulic fracture in a reservoir simulation model is required to understand the relationship between MS, fracture geometry, fracture conductivity, and production. This important step provides the "link" between MS behavior and well performance.

Most misapplications of microseismic data are associated with the wrong assumption that the microseismic events can be directly correlated to hydraulic fracture geometry and well productivity. Microseismic images are excellent indicators of hydraulic fracture location, but current processing and interpretations cannot provide reliable estimates of complex fracture geometry or the distribution of conductivity within the fracture, except when integrated with other data, interpretations and simulations.

The integration of microseismic mapping, fracture modeling and reservoir simulation is required to estimate the effective stimulated volume. Although microseismic stimulated volume may be correlated to well productivity in limited areas of some unconventional plays, variations in fracture treatment design, rock properties, and stress regime can result in significant differences in the effective stimulated volume for the same microseismic volume. The complexity of the relationship between microseismic stimulated volume and effective stimulated volume (SV) precludes simple correlations in most cases. One of the most common misapplications of microseismic interpretations is the assumption that larger SV will automatically result in increased well productivity. In areas where un-propped fracture conductivity is relatively large with respect to the matrix permeability, the effective stimulated volume may approach the stimulated volume. Characterizing un-propped fracture conductivity will be a critical factor when evaluating well performance and estimating drainage area and hydrocarbon recovery.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Nomenclature

    Nomenclature
    AbbreviationExpansion 
  • AVO

    Amplitude Versus Offset

  •  
  • bbl

    barrels, L3

  •  
  • BCFGE

    billion standard cubic feet of gas equivalent, L3

  •  
  • bpm

    barrels per minute, L3/t

  •  
  • DFN

    Discrete Fracture Network

  •  
  • ESV

    Estimated Stimulated Volume, L3

  •  
  • FC

    Fracture Conductivity (md-ft)

  •  
  • FCI

    Fracture complexity index

  •  
  • G&G

    = geological and geophysical

  •  
  • Gal

    = gallons, L3

  •  
  • ISIP

    = Instantaneous shut-in pressure, F/ L2

  •  
  • k

    permeability

  •  
  • lb

    pounds, M

  •  
  • md

    = 10-3 Darcy, L2

  •  
  • Mcf/D

    1000 standard cubic feet per day, L3/t

  •  
  • MMCF, MMscf

    = million standard cubic feet, L3

  •  
  • Mscf/d, MCFD

    1000 standard cubic feet per day, L3/t

  •  
  • MS, MSM

    microseismic, microseismic mapping, microseismic monitoring

  •  
  • MTI

    Moment Tensor Inversion

  •  
  • nd

    10-9 Darcy, L2

  •  
  • p

    pressure, F/ L2

  •  
  • PR

    Poisson’s ratio

  •  
  • QC

    Quality Control

  •  
  • SNR

    Signal to Noise Ratio

  •  
  • SV

    Stimulated Volume, L3

  •  
  • SRV

    Stimulated Reservoir Volume, L3

  •  
  • UFM

    Unconventional Fracture Model

  •  
  • σh

    minimum horizontal stress, F/L2

  •  
  • σH

    maximum horizontal stress, F/L2

SI Metric Conversion Factors

bbl × 1.589 874 e-01 = m3 
ft × 3.048 e-01 = m 
psi × 6.894 757 e+00 = kPa 
bbl × 1.589 874 e-01 = m3 
ft × 3.048 e-01 = m 
psi × 6.894 757 e+00 = kPa 

Acknowledgements

The authors thank Schlumberger for supporting this work and publication of this paper.

1
Baihly
,
J.
,
Grant
,
D.
,
Fan
.
L.
, and
Bodwadkar
,
S
.
2009
.
Horizontal Wells in Tight Gas Sands - A Method for Risk Management ToMaximize Success
.
SPE Production & Operations Journal, May
2009
,
p277
288
.
2
Cipolla
,
C.L.
,
Warpinski
,
N.R.
,
Mayerhofer
,
M.J.
, and
Lolon
,
E. P
.
2008a
.
The Relationship between Fracture Complexity, ReservoirProperties, and Fracture Treatment Design
.
Paper SPE 115769 presented at the 2008 SPE Annual Technical Conference and Exhibition
,
Denver, Colorado, USA
,
21-24 September
.
3
Cipolla
,
C.R.
,
Warpinski
,
N.R.
,
Mayerhofer
,
M.J
.
2008b
.
Hydraulic Fracture Complexity: Diagnosis, Remediation, and Exploitation
.
PaperSPE 115771 presented at the 2008 Asia Pacific Oil & Gas Conference
,
Perth, Australia
,
October 20-22
4
Cipolla
,
C.L.
,
Williams
,
M.J.
,
Weng
,
X.
,
Mack
,
M.
, and
Maxwell
,
S
.
2010
.
Hydraulic Fracture Monitoring to Reservoir Simulation: Maximizing Value
.
Paper SPE 133877 presented at the SPE Technical Conference and Exhibition held in
Florence, Italy
,
September 19-22
.
5
Cipolla
,
C.
,
Weng
,
X.
,
Mack
,
M.
,
Ganguly
,
U.
,
Gu
,
H.
,
Kresse
,
0.
,
Cohen
,
C.
, and
Wu
,
R.
2011a
.
Integrating Microseismic Mapping and Complex Fracture Modeling to Characterize Fracture Complexity.
Paper 140185 presented at the SPE Hydraulic Fracturing Technology Conference and Exhibition held in
The Woodlands, Texas, USA
,
24-26 January
.
6
Cipolla
,
C.
,
Maxwell
,
S.
,
Mack
,
M
, and
Downie
,
R
.
2011b
.
A Practical Guide to Interpretation Microseismic Measurements.
SPE 144067presented at the SPE North American Unconventional Gas Conference and Exhibition
,
The Woodlands, Texas, USA
,
14-16 June 2011
.
10
Downie
,
R. C.
,
Kronenberger
,
E.
and
Maxwell
,
S.C
.
2010
.
Using Microseismic Source Parameters to Evaluate the Influence of Faults onFracture Treatments - A Geophysical Approach to Interpretation
.
Paper SPE 134772, presented at the SPE Annual Technical Conference and Exhibition
,
Florence, Italy
,
19-22 September
.
11
Ejofodomi
,
E.
,
Bizzel
,
K.
,
Long
,
T.
, and
Mills
,
D.
,
Yates
,
M.
,
Downie
,
R.
,
Itibrout
,
T.
, and
Catoi
,
A
.
2010
.
Improving Well Completion viaReal-Time Microseismic Monitoring: A West Texas Case Study
.
Presented at the SPE Tight Gas Completions Conference held in
San Antonio, Texas, USA
,
2-3 November 2010
.
12
Fisher
,
M.K.
,
Davidson
,
B.M.
,
Goodwin
,
A.K.
,
Fielder
,
E.O.
,
Buckler
,
W.S.
, and
Steinberger
,
N.P
.
2002
.
Integrating Fracture MappingTechnologies to Optimize Stimulations in the Barnett Shale
.
Paper SPE 77441 presented at the 2002 SPE Annual Technical Conference and Exhibition
,
San Antonio, Texas, USA
,
September 29-October 2.
13
Fisher
,
M.K.
,
Heinze
,
IR.
,
Harris
,
C.D.
,
Davidson
,
B.M.
,
Wright
,
C.A.
, and
Dunn
,
K.P
.
2004
.
Optimizing Horizontal Completion Techniques in the Barnett Shale Using Microseismic Fracture Mapping
,
paper SPE 90051 presented at the SPE Annual Technical Conference and Exhibition
,
Houston, Texas, USA
,
26-29 September
.
14
Fredd
,
C.N.
,
McConnell
,
S.B.
,
Boney
,
C.L.
, and
England
,
K.W
.
2001
. Experimental
Study of Fracture Conductivity for Water-Fracturingand Conventional Fracturing Applications
.
SPE Journal
Vol.
6
No.
3
pp
288
298
(
September
2001
).
7
Cipolla
,
C.
,
Weng
,
X.
,
Onda
,
H.
,
Nadaraja
,
T.
,
Ganguly
,
U.
, and
Malpani
,
R
.
2011c
.
New Algorithms and Integrated Workflow for Tight Gas and Shale Completions
.
SPE 146872 presented at the SPE Annual Technical Conference and Exhibition
,
Denver, Colorado
,
30 October-2 November 2011.
8
Cipolla
,
C.
,
Fitzpatrick
,
T.
,
Williams
,
M.
, and
Ganguly
,
U
.
2011d
.
Seismic to Simulation for Unconventional Reservoirs
.
SPE 146876 presented at the SPE Reservoir Characterization and Simulation Conference and Exhibition held in
Abu Dhabi, UAE
,
9-11 October 2011
.
9
Daniels
,
J.
,
Waters
,
G.
,
LeCalvez
,
J.
,
Lassek
,
J.
, and
Bentley
,
D
.
2007
.
Contacting More of the Barnett Shale Through an Integration of Real-Time Microseismic Monitoring, Petrophysics, and Hydraulic Fracture Design
.
Paper SPE 110562 presented at the 2007 SPE Annual Technical Conference and Exhibition
,
Anaheim, California, USA
,
October 12-14
.
15
Inamdar
,
A.
,
Malpani
,
R.
,
Atwood
,
K.
,
Brook
,
K.
,
Erwemi
,
A.
,
Ogunda
,
T.
, and
Purcell
,
D
.
2010.
Evaluation of Stimulation TechniquesUsing Microseismic Mapping in the Eagle Ford Shale
.
SPE 136873 presented at the SPE Tight Gas Completions Conference
,
San Antonio, Texas, USA
,
2-3 November 2010
.
16
Kim
,
A
.,
2011
,
Uncertainties In Full Waveform Moment Tensor Inversion Due To Limited Microseismic Monitoring Array Geometry
,
SEG Expanded Abstracts 30, 1509 (2011)
, DOI:
17
King
,
G.E.
,
Haile
,
L.
,
Shuss
,
J.
, and
Dobkins
,
T.A
.
2008
.
Increasing Fracture Path Complexity and Controlling Downward Fracture Growth in the Barnett Shale
.
Paper SPE 119896 presented at the 2008 SPE Shale Gas Production Conference
,
Fort Worth, Texas, USA
,
16-18 November
.
18
King
,
G.E
.
2010
.
Thirty Years of Gas Shale Fracturing: What Have We Learned?
Paper SPE 133456 presented at the SPE Annual Technical Conference and Exhibition
,
Florence, Italy
,
19-22 September
.
19
Leaney
,
S.
, and
Chapman
,
C
.,
2010
.
Microseismic Sources in Anisotropic Media
.
Presented at the EAGE Conference and Exhibition
,
Barcelona, Spain
,
14-17 June 2010
.
20
Maxwell
,
S.C.
,
Urbancic
,
T.
, and
McLellan
,
P
.,
2003a
,
Assessing the Feasibility of Reservoir Monitoring Using Induced Seismicity
,
presented at EAGE Conference and Exhibition
,
Stavenger, Norway
,
2-5 June 2003
.
21
Maxwell
,
S.C.
,
Urbancic
,
T.
,
Prince
,
M.
, and
Demerling
,
C
.,
2003b
,
Passive Imaging of Seismic Deformation Associated With SteamInjection in Western Canada
,
Paper SPE 84572, presented at the SPE Annual Technical Conference and Exhibition
,
Denver, Colorado
,
5-8 October 2003
.
22
Maxwell
,
S.C.
,
Waltman
,
C.K.
,
Warpinski
,
N.R.
,
Mayerhofer
,
M.J.
, and
Boroumand
,
N
.
2006
.
Imaging Seismic Deformation Associatedwith Hydraulic Fracture Complexity
,
SPE 102801 Annual Technical Conference and Exhibition
,
San Antonio, Texas, U.S.A.
,
24-27 September
.
23
Maxwell
,
S.C.
,
Shemeta
,
J.
,
Campbell
,
E.
, and
Quirk
.
D
.
2008
.
Microseismic Deformation Rate Monitoring. Paper SPE 116596
,
presentedat the SPE Annual Technical Conference and Exhibition
,
Denver, Colorado, USA
,
21-24 September
.
24
Maxwell
,
S.C.
,
Jones
,
M.
,
Parker
,
R.
,
Miong
,
S.
,
Leaney
,
S.
,
Dorval
,
D.
,
D’Amico
,
D.
,
Logel
,
J.
,
Anderson
,
E.
,
Hammermaster
,
K
.
2009
.
Fault Activation During Hydraulic Fracturing
.
SEG Expanded Abstracts 28, 1552 (2009)
, DOI:
25
Maxwell
,
S.C.
,
Pope
,
T.
,
Cipolla
,
C.
,
Mack
,
M.
,
Trimbitasu
,
L.
,
Norton
,
M.
, and
Leonard
,
J
.
2011a
.
Understanding Hydraulic Fracture Variability Through Integrating Microseismicity and Seismic Reservoir Characterization
.
SPE 144207 presented at the SPE North American Unconventional Gas Conference and Exhibition
,
The Woodlands, Texas, USA
,
14-16 June 2011
.
26
Maxwell
,
S.C
.,
2011b
,
Microseismic Network Design: Estimating the Number of Detected Microseismic Events During Hydraulic Fracturing
,
SEG Expanded Abstracts 30, 4404 (2011)
, DOI:
27.
Maxwell
and
Cipolla
.
2011c
.
What Does Microseismicity Tell Us About Hydraulic Fracturing?
SPE 146932 presented at the SPE AnnualTechnical Conference and Exhibition
,
Denver, Colorado, USA
,
30 October-2 November 2011
.
28
Mayerhofer
,
M.J.
,
Lolon
,
E.P.
,
Warpinski
,
N.R.
,
Cipolla
,
C.L.
,
Walser
,
D.
,
Rightmire
,
C.M
.
2008
.
What is Stimulated Reservoir Volume (SRV)?
SPE 119890 presented at the 2008 SPE Shale Gas Production Conference
,
Fort Worth, Texas, U.S.A.
,
16-18 November 2008
.
29
Miller
,
C.
,
Waters
,
G.
, and
Rylander
,
E
.
2011
.
Evaluation of Production Log Data from Horizontal Wells Drilled in Organic Shales
.
SPE144326 presented at the SPE Americas Unconventional Gas Conference
,
The Woodlands, Texas, USA
,
14-16 June 2011.
30.
Moschovidis
,
Z.
et al.2000.
W.L
.
1980
.
The Mounds Drill-Cuttings Injection Field Experiment: Final Results and Conclusions
.
SPE 59115 presented at the 2000 IADC/SPE Drilling Conference
,
New Orleans, Louisiana
,
23-25 February 2000
.
31
Moos
,
D.
,
Vassilellis
,
G.
,
Cade
,
R.
,
Franquet
,
J.
,
Lacazettel
,
A.
,
Bourtembourg
,
E.
, and
Daniel
,
G
.
2011
.
SPE 145849 presented at SPEAnnual Technical Conference and Exhibition
,
30 October-2 November 2011
,
Denver, Colorado, USA
.
32
Nagel
,
N.
and
Sanchez-Nagel
,
M.
2011
.
Stress Shadowing and Microseismic Events: A Numerical Evaluation
.
SPE 147363 presented atSPE Annual Technical Conference and Exhibition
,
30 October-2 November 2011
,
Denver, Colorado, USA
.
33
Peterman
,
F.
,
McCarley
,
D.E.
,
Tanner
,
K.V.
,
Calvez
,
J.H.
,
Grant
,
W.D.
,
Hals
,
C.F.
,
Bennett
,
L.
, and
Palacio
,
J.C
.
2005
.
Hydraulic FractureMonitoring as a Tool to Improve Reservoir Management
.
Paper SPE 94048 presented at the SPE Production and Operations Symposium
,
Oklahoma City, Oklahoma
,
17-19 April.
34
Prince
,
M.
,
Baig
,
A.
,
Urbancic
,
T.
,
Buckingham
,
K.
, and
Guest
,
A
.
2011
.
Using Moment Tensors To Determine the Effectiveness of Hydraulic Fracture Stage Spacing
.
SPE 144190 presented at the SPE North American Unconventional Gas Conference and Exhibition
,
The Woodlands, Texas, USA
,
14-16 June
2011
.
35
Rich
,
J.P.
and
Ammerman
,
M
.
2010
.
Unconventional Geophysics for Unconventional Plays
.
Paper SPE 131779 presented at the Unconventional Gas Conference
,
Pittsburgh, Pennsylvania, USA
,
February 23-25
.
36
Rutledge
,
J.T.
,
Phillips
,
W.S
, and
Mayerhofer
,
M.J
.,
2004
,
Faulting Induced by Forced Fluid Injection and Fluid Flow Forced by Faulting:An Interpretation of Hydraulic-Fracture Microseismicity
,
Carthage Cotton Valley Gas Field, Texas, Bulletin of Seismological Society of America
,
94
5
:
1817
.
37
Waters
,
G.
,
Dean
,
B.
,
Downie
,
R.
,
Kerrihard
,
K.
,
Austbo
,
L.
, and
McPherson
,
B
.
2009a
.
Simultaneous Hydraulic Fracturing of AdjacentHorizontal Wells in the Woodford Shale
.
Paper SPE119635 presented at the SPE Hydraulic Fracturing Tech. Conference.
,
The Woodlands, Texas, USA
,
19-21 January
.
39
Waters
,
G.
,
Ramakrishnan
,
H.
,
Daniels
,
J.
,
Bentley
,
D.
,
Belhadi
,
J.
, and
Ammerman
,
M
.
2009b
.
Utilization of Real Time MicroseismicMonitoring and Hydraulic Fracture Diversion Technology in the Completion of Barnett Shale Horizontal Wells
.
Paper OTC 20268 presented at the Offshore Technology Conference held in Houston
,
Texas, USA
,
4-7 May
2009
.
40
Warpinski
,
N.R.
,
Branagan
,
P.T.
,
Peterson
,
R.E.
,
Wolhart
,
S.L.
and
Uhl
,
J.E
.
1998
.
Mapping Hydraulic Fracture Growth and Geometry Using Microseismic Events Detected by a Wireline Retrievable Accelerometer Array
.
SPE 40014,1998 Gas Technology Symposium
,
Calgary, Alberta, Canada
,
March 15-18
.
41
Warpinski
,
N.R.
2009
.
Integrating Microseismic Monitoring With Well Completions, Reservoir Behavior, and Rock Mechanics
.
Paper SPE125239 presented at the Tight Gas Completions Conference
,
San Antonio, Texas
,
June 15-17
.
42
Weng
,
X.
,
Kresse
,
0.
,
Cohen
,
C.
,
Wu
,
R.
and
Gu
,
H.
2011
.
Modeling of Hydraulic Fracture Network Propagation in a Naturally FracturedFormation
.
Paper SPE 140253 presented at the SPE Hydraulic Fracturing Technology Conference and Exhibition held in
The Woodlands, Texas, USA
,
24-26 January
2011
.