Multi-client seismic data plays a significant role in the development of onshore unconventional assets for oil and gas companies. The multi-client model is beneficial because oil and gas companies can access spatial subsurface data at a much lower price than comparable proprietary data.
Traditionally, these data sets were used to interpret structures to ensure that horizontal wellbores landed in the desired target interval. The use of seismic data as a decision support and risk reduction tool in unconventional plays continues to grow; its applications go beyond landing in the reservoir and staying in the reservoir. Now, seismic data is used as part of an integrated approach to estimate rock properties and even the geomechanical behaviour of the reservoir. This enables a greater understanding of how to optimise development strategies over available acreage. The accuracy of these estimates is critical to optimising well spacing and mitigating ‘parent-child’ well interference. To increase the accuracy of these estimates between wells, high-fold high-resolution seismic data is required. Furthermore, to unlock the predictive potential of this data, careful amplitude variation with offset (AVO)-compliant seismic imaging and elastic rock property estimation are also needed.
Seismic trace density
Recently, onshore seismic data – particularly in the Permian Basin, US – has experienced a paradigm shift in terms of seismic trace density. Fairfield Geotechnologies commenced its Delaware Basin, US, multi-client programme in 2012. The first survey acquired in the area had nominal fold of approximately 300 within 82.5 ft x 82.5 ft bins, which was considered a dense seismic survey at the time. C-Ranch/Mud City is the company’s most recent multi-client acquisition. It is located over the transition zone from the Central Basin Platform to the Midland Basin and has a nominal fold of approximately 1000 within 41.25 ft x 41.25 ft bins (Figure 1).
Figure 1. DB1 fold plot vs C-Ranch/Mud City fold plot.
Because bin sizes of this new acquisition have dropped from 82.5 ft x 82.5 ft down to 41.25 ft x 41.25 ft, the horizontal resolution is effectively quadrupled (Figure 2). The trace density associated with C-Ranch/Mud City is over 10 times the number of traces per square mile when compared to other modern onshore seismic surveys. New surveys also benefit from better utilisation of seismic sources, and vibroseis sweeps are staying lower for longer. Low dwell sweeps starting at 2 Hz provide onshore broadband data and benefit the vertical resolution of the seismic data.
Figure 2. Bin and fold comparison showing greater horizontal resolution and higher fold.
In summary, the interpretation of finer-scale geologic features can be achieved for the following reasons:
- Dense seismic acquisition increases horizontal resolution.
- Low dwell vibroseis sweeps increase vertical resolution by extending bandwidth on the low end of the frequency spectrum.
- High trace density extends useful bandwidth by lowering noise content of the stacked section.
The increase in fold and horizontal resolution dramatically reduces the acquisition footprint in the shallow portion of the seismic image. Another benefit of seismic acquisition at C-Ranch/Mud City is that the fold in both near and far offsets has increased significantly when compared to legacy data. While legacy seismic surveys seem to have adequate fold in the midrange of offsets, the near and far offsets are not adequately sampled. Improvements in near and far offsets allow subsequent inversion of the seismic data to be much more accurate and hence more predictive between wells. Furthermore, the improved acquisition design provides more azimuthal contributions out to offsets in excess of 3 miles. By acquiring full azimuth data, the measurements of anisotropic attributes from velocity and amplitude analysis are more accurate.
Processing seismic data
To visualise and resolve fine-scale geologic features, the seismic data must be carefully processed. AVO-compliant seismic imaging workflows are required to supplement dense modern-day seismic surveys. Generally, these surveys are acquired using geophones with 10 Hz low-cut filters. However, in C-Ranch/Mud City, for example, the vibroseis sweeps start at 2 Hz. A low frequency geophone correction is used to recognise the benefit of these low frequencies (Figure 3).
Figure 3. 2-10 Hz filter panel from Quail Ridge East multi-client survey, showing before and after geophone correction.
Better sampling of surface-wave fields enables processing tools to more accurately model and remove unwanted noise such as ground-roll (Figure 4). This is because the surface noise is not aliased as it is in legacy surveys. True amplitudes are achieved by using the latest surface-consistent processing technology. Furthermore, careful attention is given to the phase of the seismic data at key stages in the imaging sequence, particularly deconvolution. Even dense seismic acquisition cannot overcome irregularity in sampling caused by a variety of surface cultures.
Figure 4. Before and after ground roll removal on C-Ranch/Mud City data.
However, 5D interpolation techniques can be used to mitigate the effects of this irregularity. The 5D interpolation algorithm constructs pseudo-data to fill in gaps in acquisition by using the surrounding data. Once again, this process benefits from dense seismic acquisition because the 5D interpolation algorithm is much better informed by the surrounding data. This leads to a better signal-to-noise ratio, true amplitudes, and consistent phase, which in turn generates a more stable wavelet. By comparison, traditional seismic processing on legacy surveys required the use of heavy-handed processing tools, which resulted in more wavelet variation across the survey and less low frequency content. With the improved wavelet stability and signal-to-noise ratio, imaging is able to resolve much more subtle features inherent in the data. The difference between legacy processing techniques and modern practices can be seen in Figure 5.
Figure 5. Legacy vs reprocessed seismic section.
Optimising pre-stack data
A primary goal of the imaging sequence is to optimise the pre-stack data for use in seismic inversion. The imaging technologies applied benefit seismic inversion by providing true amplitude, broad-bandwidth, high signal-to-noise ratio seismic data for input. High-quality seismic data obtained from dense acquisition and AVO-compliant imaging facilitates the estimation of high-quality attributes. On its own, seismic data represents a relative measurement of reflectivity related to the interfaces of subsurface impedance contrasts. However, simultaneous inversion of seismic data predicts elastic rock properties related to geologic layers. Because the data is band-limited, the low frequency content absent from it needs to be modelled to arrive at an absolute estimate of elastic rock properties such as acoustic impedance, Poisson’s ratio, or Young’s modulus. By ensuring that the bandwidth is extended to the lowest frequencies possible through acquisition and imaging, inversion provides a much more seismic data-driven result that is also more predictive than similar approaches applied to legacy seismic datasets. Traditionally, low frequencies absent in the seismic data are modelled by interpolating well logs that have been frequency-filtered to fill this void across geologic boundaries throughout the seismic survey. While this is a valid and widely used approach, the final results are biased at the well locations used to define the low frequency model. Recent advances in seismic inversion reduce the bias of well control by iteratively deriving low frequency information while constraining the model based on seismic facies. This provides even greater seismic data-driven predictability between wells by reducing low frequency bias. The result of such an approach is a better estimate of elastic rock properties.
Rock properties acquired from seismic facies are used to better visualise and understand the reservoir between wells. They also enable the interpretation of lateral heterogeneity between wells. The temporal and lateral changes identified in the reservoir are used to plan better wells and optimise asset development. This not only requires the identification of the top and base of the reservoir, but also the interpretation of heterogeneities found within the reservoir. In the Delaware Basin, siliciclastic seismic facies are indicative of productive reservoir rock while carbonate facies are not. The interpretation of seismic characterisation studies conducted by the company in the northeastern portion of the basin highlight these heterogeneities (Figure 6).
Figure 6. Red tank arbitrary line of seismic facies.
For example, the first Bone Spring sand is a Leonardian-aged siliciclastic unit in the Delaware Basin stacked play. Within this sand a significant carbonate debris flow was interpreted that could have had detrimental effects on horizontal drilling in that zone if it was not interpreted ahead of the drill bit (Figure 7). Furthermore, inversion provides estimates of rock properties such as Poisson’s ratio and Young’s modulus, which are important properties to consider for geomechanical applications. A primary application of geomechanical analysis is the estimation of the maximum and minimum horizontal stress field in the reservoir. This can be used to properly orient horizontal wells and model how a given frac design will propagate through the rock. Lateral wells are often drilled and completed perpendicular to the maximum horizontal stress direction to maximise lateral fracture propagation, which can benefit production.
Figure 7. Interpretation fo carbonate debris flow within the first Bone Spring sand.
Planning and drilling wells in good reservoir rock gives oil and gas companies the best chance at achieving positive returns on their operational investments. This is critical in an era where free cash flow drives operations. New wells need to generate a positive internal rate of return to fund operations moving forward. Clearly, modern multi-client seismic data has an important role in achieving this goal. The multi-client model makes dense seismic data affordable for oil and gas companies, where on its own it might be cost-prohibitive. Forward-looking oil and gas companies developing unconventional assets have the potential to generate significant value by integrating modern, dense seismic data into their decision-making processes.
Read the article online at: https://www.oilfieldtechnology.com/special-reports/06072020/the-shifting-sands-of-seismic/