Geotrace Technologies, Inc.
Search
Products / Services
Spacer
Reservoir Seismic
Reservoir Services
Data Integration Services
Case Studies / Technical Talks
Technical Flyers

Case Studies/Technical Talks

Geotrace's technical capabilities are often showcased in case studies, presentations at trade shows or regional meetings. The most recent technical papers can be accessed below; earlier information appears in the Archives. Adobe Acrobat Reader may be required.

Reservoir Services
pdfA 2-Stage Approach to Broadband Processing for Improved Stratigraphic Interpretation in the Sergipe Basin, Brazil© (presented at SBGf, Rio de Janeiro, Brazil, August 3-6, 2015)
The narrow frequency bandwidth found in conventionally acquired marine seismic is due in part to the source and receiver ghosts, which lengthen the source wavelet. Additionally, as seismic energy travels through the earth both low and high frequencies are attenuated, though by different amounts. Recent advances in processing technology allow for conventionally acquired seismic data to benefit from increased bandwidth and corrected ghost wavefields, giving more usable high and low frequencies. This extension in bandwidth can lead to seismic images that reveal more detailed geologic events. We present a 2-stage methodology to broadband seismic processing which includes a pre-stack de-ghosting process and a post-stack bandwidth recovery process.
Optimizing reservoir characterization through seismic data conditioning: A case study Block 22/NCMA4, Trinidad© (presented at the Caribbean Geological Conference, Port of Spain, Trinidad & Tobago, May, 2015)

The oil and gas industry is constantly looking at ways to bridge the gap between seismic and rock properties. Seismic data is now routinely adopted for quantitative reservoir characterization of rock properties, however the quality of the input data is critical to providing robust AVO attributes. Revisiting legacy towed streamer seismic data and their products for reservoir characterization have become increasingly common. Given the high cost associated with acquiring new data, companies are constantly looking at ways to improve the fidelity of existing seismic datasets via reprocessing. A case study over 450km2 in Block 22 and NCMA-4, North Coast Trinidad is presented using a 2011 vintage towed streamer dataset.

Risk analysis using bootstrap in combination with inversion© (presented at EAGE, Amsterdam, The Netherlands, June 16-19, 2014)

We show in this paper a new approach to statistical risk analysis using the bootstrap method in conjunction with model based inversion. It is a posteriori approach combining the results of deterministic seismic inversion and rock physics analysis. It’s an effective and efficient way to manage drilling risk and uncertainty. We show an example from the Texas Gulf Coast and demonstrate its effectiveness.

Layer stripping for azimuthal parameter determination in unconventional plays© (presented at EAGE, London, UK, June 10-13, 2013)

The correct imaging of Fractured Reservoirs common in many unconventional plays involves the extension of the migration algorithms to include azimuthal anisotropic (HTI) corrections. More often than not vertical anisotropy (VTI) is also present demanding the extension of the algorithms to an orthorhombic anisotropy. The set of anisotropic parameters derived in our original work (Wojslaw & Stein 2010) had a shortcoming. It overlooked the effect of the overburden, effectively describing RMS-like anisotropy and projecting (or foot-printing) into the deeper layers the effects of the shallow ones. This produces an overestimation of the anisotropy values and an incorrect fracture orientation. This paper will describe the layer stripping methodology and its applications to a real exploration situation that removes the effect of the overburden.

Can high resolution seismic data improve pore pressure estimation and interpretation?© (presented at EAGE, London, UK, June 10-13, 2013)

The deepwater exploration and production business faces many challenges from understanding the petroleum system including migration, reservoir, and seal in a basin to ultimately drill a well, make a discovery, and produce from a field. One of the many critical components in the oil and gas production chain is the risk of encountering and mitigating abnormal pore pressure problems above normal hydrostatic pressure during the drilling phase without losing a well or damaging a reservoir formation. First we explain a holistic multi-dimensional (nD) high resolution imaging technique to improve the seismic image and seismic velocity resolution in time and depth. Then we describe a calibrated pressure model which transforms higher resolution seismic velocity into high resolution pore pressure and other pressure attributes. Finally we examine the impact and benefit of high resolution for pressure estimation and quantitative interpretation, well design and drilling plan, and an uncertainty assessment. The case studies demonstrate that high resolution seismic and velocity data lead to high resolution pressure attributes, more reliable quantitative pressure interpretation, and better drilling risk management.

Shale play characteristics and correlation with anisotropic logs - a case study of Eagle Ford Shale in Texas© (presented at EAGE, London, UK, June 10-13, 2013)

We used a model based inversion solving for Vp, Vs and Density and computed Young's modulus and Poisson's ratio to determine rock brittleness/ductility and total organic carbon (TOC) from rock properties. Although the lower Eagle Ford shale appears fairly consistent in its properties, we note specific variations in those rock properties that suggest they can be used to improve prospecting for hydrocarbons in the Eagle Ford shale play.



Reservoir Seismic
pdfImproved Subsurface Imaging and Interpretability through Broadband Reprocessing of Legacy Seismic Data: Examples from North West Shelf Australia© (presented at ASEG-PESA-AIG 2016, Adelaide, South Australia, 21-24 August 2016)

The reprocessing of legacy seismic data can be a time and cost effective means of obtaining an improved image of the subsurface, particularly when compared to the acquisition of new seismic data. The investment that has been made over the years in acquiring the many thousands of kilometres of seismic data offshore Australia has been preserved by Geoscience Australia, which houses an extensive collection of petroleum data including seismic survey data. Much of this data is available to the petroleum industry for reprocessing, facilitating the potential to enhance the data’s value for regional reconnaissance and interpretation.

Two marine examples are shown from North West Shelf Australia where reprocessing was performed on seismic data from two different surveys acquired in 1993. The first example is from the Northern Carnarvon Basin, and the second example is from the Browse Basin. These two examples demonstrate how the uplift attained from a modern broadband processing flow can yield a vastly improved subsurface image, which in turn can assist with interpretation. The reprocessing workflow (which was similar for both surveys) is discussed, as well as some insights into how the improved data benefit the interpretation and understanding of subsurface geology.

pdf3D CRS Parameter Estimation with Simultaneous Multi-dimensional Local Slopes© (presented at EAGE, Madrid, Spain, 1-4 June 2015)
The Common Reflection Surface (CRS) is essentially a generalization of the well-known Common Midpoint (CMP) stack. The full 3D CRS stack requires eight parameters, versus only three in 2D. The method of parameter determination is generally a coherence analysis based on semblance and applied to a full search of the parameter space. While expensive in 2D, it can become prohibitive in 3D. One way to lessen the computational load is to scan for the parameters separately, while holding the others fixed, although the solution is no longer optimal. We propose instead to consider data that has already been moved out, and recast the 3D CRS stacking operator in terms of residual velocity, dip, and curvature. The CRS parameters in the new formulation form a (hyper-)plane through the time slices. The coefficients of this plane are determined by multiple linear regression, and the eight CRS parameters are calculated from local slopes with a vector plane-wave destruction technique. This scheme maintains the optimal results of global simultaneous estimation while avoiding the long runtimes of semblance-based dip search methods.
pdfDeghosting Ultra High Resolution 3D Geo Hazard Data© (presented at EAGE, Madrid, Spain, 1-4 June 2015)
The use of P-cable, very accurate navigation and GPS positioning techniques have produced an Ultra High Resolution 3D data set called SAFE-BAND with frequencies in the kilohertz range. The resolution achievable at these frequencies is on the order of a few feet vertically and horizontally. One of the most important pieces of information that can be extracted from this data set is a very finely detailed water bottom. Ghost generation and recording by submerged sources and receivers produce an interference pattern that makes data interpretation ambiguous and many times impossible. One of the challenges in deghosting data is the non-stationary nature of seismic. To address this issue and be able to derive a time varying filter we have chosen to work in the continuous wavelet domain (CWT). The CWT provides a solid mathematical framework for dealing with the non-stationary nature of seismic data, by allowing a time variant analysis for the filter design and application. The results show that this technique is effective for deghosting this very high fidelity data and is able to produce an accurate water bottom, unmasking subtle topographical features in the water bottom as well as geological features in the near surface geology.
pdfSingle Pass Multilayer Determination of Anisotropic Velocity Models with Well Constraints© (presented at EAGE, Madrid, Spain, 1-4 June 2015)
Prestack. depth migration of compressional data in the anisotropic case requires three scalar fields: VNMO, epsilon, and delta. The short-spread moveout velocity VNMO will create flat gathers in both the isotropic and anisotropic cases, but with incorrect depths for non-zero delta, while the "moveout" part at the correct depth is controlled by epsilon. The anisotropic parameter delta can only be obtained using the additional information provided by well measurements. Epsilon and delta are generally estimated in a layer-stripping workflow, which requires a new PSDM for each anisotropic layer. We present a method for constructing the delta and epsilon models in a single pass for all layers. A key point is that the delta layers are bounded by the well depths, and not the seismic horizon depths. Multilayer epsilon models are obtained by interactive raytracing of residual moveout curves as a function of epsilon which overlay the migrated data gathers. The entire workflow consists of an isotropic migration, a single pass delta volume creation, anisotropic migration to tie the wells at near offsets, a single pass determination of the epsilon model, and a final migration that produces flat gathers which tie the wells.
pdfTrue depth anisotropy in complex geological settings© (presented at SBGf, Rio de Janeiro, Brazil, August 3-6, 2015)
Over the years, our increased understanding of anisotropy has allowed us to improve the way we treat velocity at a single location in the subsurface. In complex geological settings where positioning is important, accurately correcting for anisotropy becomes critical for resolution and accurate well placement. Simple constant anisotropic parameters, in most cases, are not an accurate representation of the earth given the variation of the heterogeneity of the subsurface. It is therefore essential for anisotropy to vary spatially and vertically. In this paper we demonstrate a method and workflow for addressing the variability of the delta and epsilon parameters across many wells and show how using our true depth technique and picking for epsilon we can greatly improve our anisotropic models and hence the resolution of the stack.
pdfA 2-Stage Approach to Broadband Processing for Improved Stratigraphic Interpretation in the Sergipe Basin, Brazil© (presented at SBGf, Rio de Janeiro, Brazil, August 3-6, 2015)
The narrow frequency bandwidth found in conventionally acquired marine seismic is due in part to the source and receiver ghosts, which lengthen the source wavelet. Additionally, as seismic energy travels through the earth both low and high frequencies are attenuated, though by different amounts. Recent advances in processing technology allow for conventionally acquired seismic data to benefit from increased bandwidth and corrected ghost wavefields, giving more usable high and low frequencies. This extension in bandwidth can lead to seismic images that reveal more detailed geologic events. We present a 2-stage methodology to broadband seismic processing which includes a pre-stack de-ghosting process and a post-stack bandwidth recovery process.
pdfSimultaneous estimation of CRS parameters with multi-dimensional local slopes© (presented at SEG, Denver, Colorado, October 26-31, 2014)
Global determination of CRS stacking parameters by semblance optimization over mult-dimensional ranges is an expensive endeavor, even for 2D seismic datasets. Direct estimation of these parameters by local slope methods is much faster, but these are typically applied in each dimension separately, not globally, compromising the quality of the resulting stack. Furthermore, the effect of a heterogeneous overburden on the velocity parameter(s) remains an issue in many implementations. Velocity analysis and moveout based on ray-traced offset-dependent velocities remove these effects, turning the CRS velocity parameter estimation into a residual velocity estimation, which can also be cast as a local slope determination. A multiple linear regression calculation of the gradients for plane wave destruction allows all parameters to be estimated simultaneously.
pdfImproved plane wave destruction for iterative local slope estimation© (presented at EAGE, Amsterdam, The Netherlands, June 16-19, 2014)
The estimation of local slopes has been an important topic for numerous seismic processing applications. Many local slope algorithms are based on the plane wave destruction approach, using the time derivative of the seismic trace along with a horizontal gradient trace to generate estimates of the dip. Differentiation increases high frequencies in the data, limiting the accuracy of these dip estimates. Alternately, the plane wave destruction equation may be recast in terms of integration. In this case the original seismic trace is used along with the causal integration of the horizontal gradient trace. Since there is no differentiation to boost high frequency noise, and since numerical integration is itself a smoothing process, this approach leads to better-behaved dip estimates. In both cases, the finite difference approximation of the gradient leads to a biased estimate. However, as long as the local event slopes are not aliased, the dip estimates will be of the correct sign, and an iterative application of the algorithm will lead to accurate local slopes.
pdfComparison of centered and traditional offset vector tiling for a wide azimuth land 3D survey© (presented at EAGE, Amsterdam, The Netherlands, June 16-19, 2014)
We present a centered offset vector tile (OVT) scheme and illustrate its benefits for wide azimuth (WAZ) land 3D seismic surveys. The OVT method is used to produce the single fold prestack data volumes that cover the full 3D survey area. All traces in a single OVT volume share similar offset and azimuth attributes. They also provide new opportunities for nD noise attenuation, nD interpolation, and nD imaging that lead to higher signal to noise ratio and better subsurface image. Traditionally, the first four tiles are built around the intersection of a cross-spread so the nearest offsets are distributed equally to these four tiles, and then the other tiles are generated around them and spread out from near to far offsets. Our centered tiling method centers the first tile at the intersection of a cross-spread and then the remaining tiles are spread out around it from near to far offsets. Centered tiling approach ensures a minimum offset OVT at the center of a cross-spread acquisition and images the shallow section better. A WAZ land 3D dataset is tested to demonstrate the difference and advantage of the centered OVT method.
pdfUnderstanding the reservoir and drilling risk with multidimensional high resolution imaging© (presented at EAGE, Amsterdam, The Netherlands, June 16-19, 2014)
Multidimensional (nD) imaging technology has made tremendous progress in improving wide azimuth seismic image quality by reducing noise interference, minimizing near-surface impact, and mitigating missing data due to acquisition irregularity. This paper first examines 1D to 5D (nD) techniques for noise attenuation and interpolation in order to achieve a better subsurface image. Then, it reviews the impact of a seismic wavelet in terms of its bandwidth and dimensionality from 1D up to 4D (nD), for high resolution (HR) imaging. Furthermore, we use several real examples to illustrate the advantage of applying a nD HR imaging methodology by combining nD imaging and HR imaging for detailed structural and stratigraphic interpretation and reservoir characterization in various lithology and play types. Finally, we detail two case studies: (1) thin sand reservoirs from lowstand deltas and slope fans in a shelf margin depositional environment and (2) tight carbonate reservoirs from distal slope wackestones, reef and bank complexes, and back reef lagoonal deposits in a progradational sequence to demonstrate the benefit of multi-dimensional imaging for reservoir characterization and drill risk management.
pdfStudying a spatially variant deghost technique for conventional streamer data© (presented at EAGE, Amsterdam, The Netherlands, June 16-19, 2014)
In this paper, we present a spatially variant deghost technique to eliminate the ghost effect and recover
the lost high frequency data for marine deep-towed streamer data acquired at a constant depth.
Besides appraising the deghost effectiveness, we validate this deghost technique by comparing the
result of a broadband-processed deep-towed streamer 2D line with a nearby shallow-towed streamer
2D line acquired at a shallower constant depth as well as evaluating the similarity and difference of
their subsurface image, seismic resolution and phase. Furthermore, the technique is tested with
premigration, prestack, and poststack applications to examine its usefulness and robustness. The
assessment of this study demonstrates its effectiveness and versatility for vintage and new marine
streamer seismic data.
pdfExtracting true anisotropic values using raytraced two parameter velocity analysis© (presented at EAGE, London, UK, June 10-13, 2013)
We present a methodology that allows the simultaneous determination of velocity and true eta in a production setting. Rather than deal with non-hyperbolic gathers by higher-order moveout equations, we take a new approach of using a different velocity function for each offset (Hellman, et al., 2012). At each analysis location, interval velocities are inverted from the near-offset rms velocities and then travel times are determined by anisotropic ray tracing through the implied model determined by the velocity and eta picks. These raytraced travel times are inserted into the hyperbolic moveout equation and rearranged to form offset dependent velocities, which are used to move out the gathers. Etas and velocities are iterated until the updated gather is flat. A primary benefit of the method is that the eta values are completely decoupled from the non-hyperbolic effect of the layered overburden, so that the derived etas are geophysical, and not "effective." The velocities and etas are not necessarily picked on the same grid, allowing the processor to control the velocities in a much finer grid than is necessary for the anisotropy. The final set of picks can then be re-gridded to be on the same grid for further processing steps.
pdfShale play characteristics and correlation with anisotropic logs - a case study of Eagle Ford Shale in Texas© (presented at EAGE, London, UK, June 10-13, 2013)
We used a model based inversion solving for Vp, Vs and Density and computed Young's modulus and Poisson's ratio to determine rock brittleness/ductility and total organic carbon (TOC) from rock properties. Although the lower Eagle Ford shale appears fairly consistent in its properties, we note specific variations in those rock properties that suggest they can be used to improve prospecting for hydrocarbons in the Eagle Ford shale play.