1
AN OVERVIEW OF SEISMIC ATTRIBUTES

Seismic attributes play a vital role in the interpretation of subsurface geological features such as faults, fractures, folds, channels, diapirs and reefs, and in inferring dynamic and static properties of subsurface reservoirs. Hence, understanding of attributes and their extraction from surface seismic data are crucial for the illumination of subsurface structures and properties. This chapter provides an overview of seismic attributes, their historical evolution, and fundamental characteristics or properties that can differentiate objects and their subsurface disposition.

1.1 Introduction

Seismic attributes have been used over the past few decades to infer subsurface properties and geologic features from reflection seismic data. These attributes illuminated geologic features, revealed structural architecture, and quantified specific physical properties. The analysis of seismic attributes plays a pivotal role in petroleum exploration through the delineation and interpretation of subsurface faults, fractures, folds, channels, diapirs, reefs (Figure 1.1), etc. These subsurface features act as traps for hydrocarbon accumulation and help in predicting dynamic and static characteristics of subsurface reservoirs (Chopra & Marfurt, 2007). Since inception, a large number of attributes have been generated from seismic data, which have been efficiently utilized to describe and delimit different geologic targets of interests.

Figure 1.1 Volumetric display of seismic cube and corresponding seismic attributes (a-b and c-d), demonstrating their efficiency in describing different subsurface geological features (after Kumar and Sain, 2018; Kumar et al., 2019; VC: Volcanic Core).

Seismic attributes are used to extract information from the pre-stack or post-stack data i.e., gathers or volumes. Pre-stack attributes treat seismic data as records of seismic reflections that are associated with the P- and S-impedances, P- and S- wave velocities, amplitude variation with offset (AVO), attenuation, anisotropy, AVO intercept, and gradients. However, post-stack attributes consider seismic data to be a representation of Earth's subsurface image that includes a large family of attributes, e.g., complex trace attributes, interval attributes, horizon-based attributes, time-frequency attributes, and waveforms (Barnes, 2016; Al-Shuhail et al., 2017). The attributes, as a whole, follow a unified characteristic, based on which maximum information about a target can be extracted and subsurface architecture of a geologic body can be inferred from data. Thus, they act as filters, designed in such a way as to illuminate the properties of interest by setting aside those which are of no interest. Though both domains divide the family of seismic attributes by assigning different means of usage, attribute analysis has received the utmost attention in image processing and enhancement, which aims to extract valuable subsurface information from surface data.

This chapter shows how the seismic attributes have evolved and are used for enhancing characteristic properties stored within the seismic data. Most of these properties form the basis for designing and formulating seismic attributes for subsurface interpretation from surface data.

1.2 Historical Evolution of Seismic Attributes

The 1920s marked the beginning of field reflection seismic experiments, which mapped subsurface geological structures by identifying reflections and converting the arrival times into corresponding depths. This practice continued through the 1950s until the late 1960s. The advent of Analog-to-Digital (A/D) conversion techniques facilitated seismic data processing on digital computers. However, this digital revolution still did not stream interpretation practices by geophysicists who could correctly map subsurface geology from data. Rummerfeld (1954) was one of those visionaries who could qualitatively use reflection characteristics to interpret subsurface stratigraphy from seismic data. This event triggered the enthusiasm of geophysicists for critically understanding the properties from seismic data for inferring subsurface geological structures. Koefoed (1955) laid down his interpretational insights into signal processing from amplitude variations with offsets (AVO), which led to the interpretation of subsurface lithological properties. Thereafter, several geophysicists (Merlini, 1960; Savit, 1960) documented their pioneering works for the interpretation of seismic reflection data. The digital revolution brought about significant changes in signal processing and quality (Yilmaz, 2001). Slowly, the interpretation of seismic reflections in inferring subsurface stratigraphy became a routine job.

The late 1960s witnessed a significant discovery from such practices. Recognition of "bright spots" from seismic data by Soviet geophysicists opened up a new era in interpretation strategies. These anomalous features aroused interest in the direct detection of hydrocarbons, geared up seismic explorationists, fascinated many seismic contractors, and led to successful exploration cases in the Gulf of Mexico. The bright spot detection through the 1970s captivated the interpretation community, making the first seismic attribute "reflection amplitude."

Reflection strength (Figure 1.2) is a classic example of amplitude attribute, designed by Anstey (1972).

Figure 1.2 Isometric display of reflection strength attribute (after Anstey, 1972).

Today this attribute is considered the most important and powerful of all existing seismic attributes. Once the amplitude phenomena led to the direct search for hydrocarbons, researchers immediately became curious about the frequency characteristic of the signal. When seismic waves propagated through a gas reservoir, it was observed that the signal suffered from higher frequency being washed out, leaving behind the lower frequency, thereby causing a shadow. Sheriff (1975) called this "Frequency Shadow Effect", which directly indicates the presence of a gas reservoir. Dobrin & Savit (1960) documented that such attenuation could be used for the quantification of rock quality factor. Anstey (1977, 2001, 2005) developed a novel procedure for attribute analysis, in which he demonstrated the use of color displays for analyzing seismic attributes. The "Complex Trace Analysis" made its first appearance in the 1976 annual meeting of the Society of Exploration Geophysicists (SEG) in the form of seminar papers by Taner, Sheriff and his group, which later on became the first masterpiece of work. Based on this concept, Taner & Sheriff (1977) and Taner et al. (1979) developed five attributes: instantaneous amplitude, instantaneous phase, instantaneous polarity, instantaneous frequency, and weighted average frequency. While these developments were gaining pace, Peter Vail and his group formulated the principles of seismic stratigraphy. Being inspired by the pioneering works of Taner and his group, the complex trace attributes found their place in explaining the seismic properties related to the stratigraphy. Several other attributes, e.g., root mean square amplitude, zero-crossing frequency, and cosine of phase, which were observed in early 1980s, were also regarded to be more comprehensible substitutes for complex trace seismic attributes. On the advent of 3D seismic data and the use of computer systems during the mid-80s, attribute maps and their analysis came to light. The first attribute maps were the simple attributes extracted from seismic amplitude volume (Denham & Nelson, 1986), horizon attributes, and horizon-guided interval attributes (Bahorich and Bridges, 1992; Dalley et al., 1989; Hoetz & Waters, 1992; Rijks & Jauffred, 1991). Bahorich & Farmer (1995) further developed 3D discontinuity attributes, which, when displayed through time and horizon slices, distinctly revealed faults, salt domes, and meandering channels. This caused curiosity among interpreters and led to the revitalization of seismic attribute analysis. This development opened an avenue for assessing other 3D properties such as dips, azimuths, curvatures, parallelism, etc. (Marfurt et al., 1999; Oliveros & Radovich, 1997; Randen et al., 2000; Taner, 2001). Mapping of thin beds and channel deposits from seismic data led to the development of spectral decomposition, which also made a breakthrough in seismic attribute analysis (Gridley & Partyka, 1997). Such an approach, coupled with tuning thickness analysis, led to a step forward for the quantitative application of seismic attributes.

Multi-attribute analysis gained importance and found a place in routine applications for seismic data interpretation. It is notable that the supervised methods were more preferred, as these methods could be trained to produce results of geological importance (Aminzadeh & de Groot, 2004; Hampson et al., 2001; Meldahl et al., 2002; Nikravesh et al., 2003). However, extraction of meaningful geology through unsupervised methods remained challenging. To date, a plethora of seismic attributes have been developed to capture the responses from subsurface geologic features for meaningful interpretation. Seismic attributes and their evolution have been very important in quenching the thirst of...