Application of super-virtual seismic refraction interferometry to enhance first arrivals: A case study from Saudi Arabia

Abstract

Application of super-virtual seismic refraction interferometry C omplex near-surface anomalies are one of the main onshore challenges facing seismic data processors. Refraction tomography is becoming a common technology to estimate an accurate near-surface velocity model. This process involves picking the first arrivals of refracted waves. One of the main challenges with refraction tomography is the low signalto-noise ratio characterizing the first-break waveform arrivals, especially for the far-offset receivers. This is especially evident in data recorded using reflection acquisition geometry. This low signal-to-noise ratio is caused by signal attenuation due to geometrical spreading of the seismic wavefield, near-surfacegenerated noise, and amplitude absorption. Super-virtual refraction interferometry improves the quality of the first-break picks by enhancing the amplitude of the refracted waves and attenuating the amplitude of the random noise.The theory of refraction interferometry was developed by Dong et al. (2006) related traces over all postcritical source positions yields a trace with a virtual refraction event that has an enhanced signal-tonoise ratio. This enhancement can be as high as , where N is the number of sources that contribute to the generation of this particular virtual head wave. They demonstrate this method on land data shot over a salt dome in Utah and later Nichols et al. (2010) showed its effectiveness in a hydrogeophysical research site in Idaho.A problem with refraction interferometry is that, if only the head wave arrivals are correlated with one another, the virtual head-wave trace has the correct moveout pattern. It has an unknown excitation time, so as a remedy, Dong et al. suggested that the source be "virtually" relocated to the surface by calibrating the virtual stacked refraction trace to an observed traveltime in the raw data. Another problem is that correlation of traces typically decreases the source-receiver offset of the virtual trace because traveltimes are subtracted and are associated with shorter raypaths. To overcome this, Bharadwaj and Schuster, Mallinson et al. (2011), andHanafy et al. (2011) presented an extension of refraction interferometry so that the receiver spread could be extended to its maximum recording extent, and the absolute arrival time can be properly accounted for. This new method creates virtual far-offset refraction arrivals by

Three-dimensional supervirtual seismic refraction interferometry

“…Generating supervirtual traces comprises of two principal stages (Alshuhail et al, 2012). The first stage involves crosscorrelation of traces recorded from postcritical sources to generate virtual traces, and the second stage is the convolution of the virtual traces with the raw traces (i.e traces before crosscorrelation) to create the supervirtual traces (Bharadwaj et al, 2012).…”

“…The first stage involves crosscorrelation of traces recorded from postcritical sources to generate virtual traces, and the second stage is the convolution of the virtual traces with the raw traces (i.e traces before crosscorrelation) to create the supervirtual traces (Bharadwaj et al, 2012). However, the process of convolving the enhanced traces with the raw traces re-introduces certain proportion of noise back into the data by generating side lobes (Alshuhail et al, 2012) around the maximum amplitude of the enhanced traces. To fix this problem, we proposed a semi-automatic first-arrival calculation method which is discussed in detail in chapter three.…”

“…It is pertinent to consider the area extent of the subsurface refractor which may not be as wide as the coverage area of the 3D acquisition design commonly used in the petroleum industry. However, if the 3D acquisition design accommodates a huge amount of sources and receivers (Alshuhail et al, 2012) in a single patch, this challenge could be overcome. Furthermore, we assume that the refracted rays follow the same ray paths in 2D supervirtual interferometry (SVI), however, this 13 concept does not apply to the 3D SVI.…”

“…This will enable the first arrival picking to be carried out automatically at better confidence level without necessarily discarding most challenging seismic data. Several studies (Alshuhail et al, 2012;Al-Shuhail et al, 2013;AlShuhail, 2015) tackled this problem, however, these studies cannot be applied directly to active 3D seismic data. Al-Shuhail et al (2013) and Al-Shuhail (2015) developed algorithms which were used successfully to enhance the first arrival of microseismic data and improve automatic picking of first arrivals of 2D seismic data, respectively.…”

“…It is pertinent to note that convolving the enhanced trace with the raw trace will re-introduce noise into the data, as a result, the initial source wavelet is no longer preserved. Side lobes are formed around the main wavelet after convolution which distorts the first arrival wave shape and poses a great challenge during picking (Alshuhail et al, 2012). To overcome this problem we consider a semi-automatic method to shift back the first-arrival on the enhanced trace to its correct time.…”