Abstract
Image logs aid in accurate fracture network characterization by providing detailed information about natural fractures in reservoirs. Moreover, larger-scale fracture mapping and detection can be accomplished by extracting seismic attributes offering important insights into the distribution of reservoir fractures. Through the extraction of seismic attributes from image logs the current study offers a novel method for locating naturally occurring fractures in a carbonate reservoir of southwest Iran. Extending the application of seismic attributes to image log data is made possible for the first time in this study by using data from electrical and ultrasonic image logs converted from DLIS to SEGY format. After converting the format of the image logs, seismic attributes were extracted and assessed for possibility of enhanced image log interpretation. To examine and compare the results of various attributes, a total of 23 structural, stratigraphic, and signal processing attributes were derived. The newly tested attributes, including the iso-frequency component, RMS amplitude, and variance (edge method), showed improved performance in identifying fractures. However, the iso-frequency component attribute was found to be very powerful in emphasizing the fractures among the different attributes examined. The findings of this study show how seismic attributes can be used to detect the natural fractures based on image logs in a reservoir resulting in more accurate fracture detection and modeling.
Introduction
The interpretation of seismic data for oil and gas exploration relies heavily on seismic attribute analysis. By analyzing several attributes, geoscientists have gained more knowledge of structures and properties in the subsurface. This means more accurate reservoir characterization and resource estimation1,2,3. The characteristics can provide clear insights into subsurface properties and can include changes in amplitude, frequency, curvature and many other elements. Geoscientists can find anomalies, faults and other features that might indicate possible hydrocarbon reservoirs by examining these characteristics4,5,6.
Image logs are highly informative when calibrated against cores and conventional logs. They provide information about lithology, sedimentary textures, paleocurrent directions, structural dip, in situ stress, and fracture patterns7,8,9,10,11,12,13,14,15,16,17,18,19. Literature review illustrates the efficiency of image logs in enhancement of sedimentological knowledge10,12,20,21,22,23, analysis of in-situ stresses24,25,26,27,28,29,30,31 and natural fractures characterization32,33,34,35,36,37,38.
Reservoir engineers use image logs to identify and analyze open fractures. This helps them better understand the reservoir’s fracture network and its impact on fluid flow and production. Also, they can evaluate the possibility of fluid flow pathways within the reservoir, characterize fracture patterns and evaluate fracture connectivity with the detailed information gained from image logs39,40. This knowledge enables engineers to optimize reservoir development strategies. It improves decisions on well placement and hydraulic fracturing design, thereby maximizing efficiency and hydrocarbon recovery. When analyzing natural fractures, the incorporation of image logs leads to more thorough reservoir characterization and facilitates well-informed decision-making when exploring and producing hydrocarbon resources41.
The image log data in this study were converted to SEGY format in order to extract the seismic attributes. High-resolution borehole image logs are essential for sedimentary reservoir interpretation. They provide millimeter-scale information on directional sedimentary structures and natural fractures17,38,42,43,44,45.
This study aims to determine whether seismic attributes applied to DLIS to SEGY-formatted converted borehole image logs can enhance the detection and visualization of natural fractures in carbonate reservoirs. The primary research question is: Can fracture characterization be improved beyond traditional visual interpretation by applying seismic attributes on high-resolution image log data? In answering this query, the study examines the comparative effectiveness of a number of attributes and identifies the potential in highlighting fracture characteristics that may otherwise be overlooked.
In this study, image logs were converted to SEGY format, enabling the application of seismic attributes directly on high-resolution borehole images. This integrated approach enhances the recognition of natural fractures and sedimentary features, enhances the accuracy of subsurface interpretation, and supports more effective geological analysis and structural event identification.
Geological setting
The studied oilfield is situated in the Dezful Embayment, part of the Zagros fold and thrust belt in southwestern Iran (Fig. 1). The Zagros region, particularly the Dezful Embayment, has exceptional petroleum potential. This is due to the existence of prolific source rocks, porous and permeable reservoirs, and good cap rocks46.
The Dezful Embayment contains a shallow petroleum system that dates from the Cretaceous to the Early Miocene and is recognized as one of the richest and most productive oil fields globally, holding approximately 8% of the world’s oil reserves47. The petroleum system comprises two source rocks, namely the Kazhdumi and Pabdeh Formations, and two reservoirs, the Asmari and Sarvak–Illam Formations. The Gachsaran and Gurpi Formations serve as caprock of the mentioned formations. Notably, the Asmari Formation accounts for 75% of Iran’s onshore hydrocarbon reserves48,49,50.
The study oilfield is a 65 km long and 4–8 km wide doubly plunging long anticline. It is famous for rich oil production from two main carbonate reservoirs: the Asmari Formation (Oligocene–Lower Miocene) and the Sarvak Formation (Middle Cretaceous). Due to tectonic-induced heterogeneities in the Asmari reservoir, the oilfield is considered one of the most geologically complex structures in the area. Severe faulting and erosion and lack of sedimentation for cretaceous succession (owing to Kharg-Mish paleo-high) led to different production conditions for each sector of the oilfield51.
A key factor influencing the reservoir’s characteristics is the natural open fracture system of the Asmari reservoir in the studied field. The permeability of the reservoir is influenced by fractures serving as channels for fluid flow. According to Saadatnia et al., 2024, the reservoir’s overall productivity in the studied field is impacted by the natural fractures influencing by the reservoir’s connectivity and storage capacity31.
The Asmari reservoir contributes nearly 85% of Iran’s total crude oil output and is globally recognized as a prominent reservoir52. Comprising cream to gray limestone, dolomitic limestone and dolomite, the Asmari Formation has a thickness of 314 m in the type section53. Serving as the cap-rock for the oilfield, the Miocene-aged Gachsaran Formation consists of seven members, including anhydrite, limestone and bituminous shale. The lower boundary of the Asmari Formation is formed by the Pabdeh Formation, a Paleocene to Eocene source rock. The location of the studied oilfield and the Cenozoic era stratigraphic column in the Dezful Embayment are depicted in Fig. 2.
The location of the studied oilfield in the Zagros fold-thrust belt.
The stratigraphic column of the Dezful Embayment.
Materials and methods
Image logs
This research utilized FMI (Fullbore Formation Micro Imager) and CBIL (Circumferential Borehole Imaging Log) data. These image logs were converted from DLIS to SEGY format for supplemental processing. There are two primary categories of image logs: electrical and acoustic (ultrasonic). Electrical image logs, such as FMI, record the electrical properties of the wellbore wall, whereas ultrasonic image logs represent the acoustic impedance.
The FMI, a pad-based electrical imaging tool, needs the pads to be securely connected to the borehole wall for data acquisition. With a vertical and azimuthal resolution of 0.5 mm, FMI image logs can point out minute features like thin laminations, stylolites, fractures, etc55. Prior to interpreting the images and identifying events, it is necessary to perform corrections, equalization and normalization on the image data. The FMI tool has four arms, each equipped with eight pads—primary pads accompanied by flappers. This allows much larger borehole coverage with 24 image buttons on each pad for a total of 192 image buttons (Fig. 3).
The major applications include fracture recognition, structural analysis and stratigraphic interpretation. Fracture recognition includes the characterization and analysis of fractures, vugs, and similar features. Structural analysis includes the determination of structural dip, evaluation of unconformities and the analysis of faults55.
Schematic of the Formation MicroImager (FMI) tool (Daws et al., 200255).
The FMI tool is a new generation tool with twice the borehole coverage of the Formation MicroScanner (FMS) tool. The extra coverage is obtained by the addition of a flapper pad below and offset from each of the regular pads. Smaller buttons are employed to increase the image resolution by approximately 20%55. The total number of buttons is increased from 64 to 192 resulting in tripling the number of samples taken (Fig. 4).
As the tool emits current into the formation, it theoretically works only in water-based mud. Mud resistivity should not exceed 50 ohm-m; however, the mud should not be too conductive. For good image quality, the ratio of formation resistivity to the mud resistivity should be below 1,000. When the mud is too conductive relative to the formation, the current tends to flow into the borehole, reducing the sharpness of the images55.
The pads, flappers and sensors of the FMI tool (Daws et al., 200255).
The CBIL enables the analysis of fractures, stress and borehole stability studies and enables structural interpretations that are uninfluenced by mud type. Through advanced borehole imaging and stress analysis applications, borehole stability and breakout information can be derived from the accurate borehole cross section as measured by the CBIL tool. For open-hole measurements and for casing internal geometry measurements in which casing resonance is not needed the CBIL tool’s transducer results in high-resolution images41.
The sonde includes a rotating transducer subassembly, which is available in different sizes to log all standard sizes of open boreholes. The proper choice of transducer subassembly is extremely important. It reduces attenuation in heavy fluids and has a high signal-to-noise ratio. This optimization improves ultrasonic pulse propagation in the borehole. The transducer is both a transmitter and a receiver, transmitting an ultrasonic pulse and receiving the reflected pulse (Fig. 5).
The physics of measurement of CBIL tool (Daws et al., 200255).
The data format conversion
The SEGY format is a widely recognized standard format for 2D and 3D seismic reflection measurement recordings, in which the subsurface data are stored. These types of files are containing essential information including trace samples, amplitudes and required details for seismic data interpretation.
Each 2D seismic SEGY file trace represents a single seismic channel and record the amplitude of seismic waves at different depths below the surface. By digitally capturing the amplitude of these waves, subsurface geological information will be provided. The SEGY file structure allows for the storage of both textual metadata, describing the file contents and binary data including the actual seismic waveforms.
The project initiated with loading a SEGY file through segyio.open() phyton function. To avoid potential problems with the files geometric properties the ignore_geometry = True parameter was used when calling the open function. Subsequently segyio. tools. metadata () was used to extract the SEGY files metadata providing an extensive overview of the file’s properties.
Subsequently, a new SEGY file was carefully generated to match the structure of the original, ensuring compatibility and ease of data transfer. In order to generate semi-seismic-readable data, it was necessary copying not just both the textual and binary information but also the header details. Additionally, the traces from the initial source file were replicated in the new file, preserving the integrity of the seismic survey data structure.
The outcome of this process was a SEGY file (*.sgy) that precisely replicates the structure and content of the original file, with a specific focus on maintaining the integrity of the seismic traces. Here is sample SEGY file header information:
DLIS file is a globally standard format, which is usually hierarchically organized, containing information such as headers, traces and frames, used to store digital well log data. Each frame can hold multiple curves, representing various measurements of desired electrical properties of the subsurface formations such as gamma ray (GR), neutron and density. This hierarchical structure enables us for efficient storage and retrieval of complex datasets. The dlis module from the dlisio package of Python software was utilized to load a DLIS file. Afterwards, the extracted curve data was initially stored in an empty list called all_curves_values. An iterable object called files provides access to the contents of the file once the load () function is called with the file path as an argument. Then, an additional loop iterates within each frame in the current logical file, within the loop iterating over each logical file (f). To obtain a group of curves for every frame the curves () method is then invoked on the frame object. To ensure that the curves array is not empty, number of curves in the current frame will be counted. The last curves length is printed if it contains data and the values of each curve are copied to the all-curves values list.
Iterating through frames and extracting curve values, enables us to navigate the hierarchical structure of DLIS files to retrieve specific pieces of information. The final output provides comprehensive details about the loaded DLIS files architecture. In particular it specifies the beginning of the DLIS datasets first logical file. The output then shows the start of the first frame in this logical file. Here is sample DLIS file information extracted to text file:
Digital well log data are stored in DLIS files, while seismic survey data are typically stored in SEGY files as a rule of thumb. This is an attempt to improve the geological interpretation by adding well log information by substituting curve values from the DLIS file for the trace data in the SEGY file. The first step of this process is to open the SEGY file in the r + mode which enables reading and writing simultaneously. In order to modify the contents of the SEGY file based on the DLIS data, this is essential for the current task. Curve values taken from the DLIS file are initialized into a temporary array called current_curve. Afterward, the code iterates through every trace in the SEGY file, adding the matching curve values from the DLIS file to current_curve. This is accomplished by obtaining the curves from the DLIS datasets (files [0]. frames [0]. curves ()) and making sure that the curve values and the traces in the SEGY file match exactly. An essential component of this procedure is handling exceptions. The try-except block in the code encapsulates the operations and is used to handle any exceptions that may arise during this process. This is necessary to protect the integrity of the data and make sure that mistakes do not result in inaccurate or corrupted data transfer. In the event that an exception is raised the code logs the problem and exits the loop stopping any more updates to the SEGY file that might lead to problems. Following the DLIS curve values being entered into current_curve the code transposes the array to conform to the expected trace data layout from the SEGY file. In the final step, the code loops through the SEGY file again. It replaces the original trace data with corresponding values from current_curve. Figure 6 shows a pseudocode for updating phase. The workflow for converting DLIS image log data format to SEGY format is also shown in Fig. 7.
Python Pseudocode for updating phase.
Workflow for converting DLIS image log data format to SEGY.
Seismic attributes
The analysis incorporated a broad set of 23 seismic attributes, strategically chosen to provide a detailed understanding of the image log features. These attributes, which are outlined in Table 1, can be classified into three groups: stratigraphic, structural and signal processing attributes.
The selection of the 23 seismic attributes employed in this research was a combination of literature justification, operational experience, and correspondence with the natural properties of image log data. The attributes were initially chosen from a broad spectrum of attribute types—seismic attributes ranging from the structural, to the stratigraphic, to the signal-based types—to allow for well-rounded analysis of the data and ensure that the various properties of fracture-related features were well captured. Secondly, all the selected attributes have a proven history of successful application in similar geophysical and geological studies. Attributes such as variance, RMS amplitude, and iso-frequency component have repeatedly proven their usefulness in outlining fractures and subsurface structural elements. Finally, the intrinsic nature of borehole image logs was a significant factor in the choice. Foremost in consideration were those properties that could emphasize low contrasts and delineate fine-scale discontinuities characteristic of natural fractures.
Sedimentary layers and other geological formations are detected and described using stratigraphic attributes. Their foundation lies in the way seismic waves reflect as they move through different kinds of rock. Geological structures like faults, folds and unconformities are identified and described by their structural characteristics. These characteristics come from the shape and direction of seismic reflections. Signal processing attributes enhance both the quality of seismic data and the interpretation of geological features.
Results and discussion
Using the petrophysical software, image logs are displayed in both static and dynamic formats for interpretation purposes. The image log in Fig. 8, from Well-A, shows a CBIL image log in this software environment. Dark-colored sinusoids in the image represent numerous natural fractures, which are apparent in two sets that are oriented perpendicularly to each other in both static and dynamic images.
The CBIL image log shows multiple natural fractures visible in static and dynamic images. These fractures are depicted as dark-colored sinusoids, with two sets observed at right angles to each other.
Figures 9, 10, 11, 12 and 13 show how the image log data were loaded into the modeling software after being converted from DLIS to SEGY format. As SEGY is the industry format for seismic data, this conversion step was essential to ensure compatibility with the modeling software. Seismic attributes were run on the SEGY image log following the loading of the image data in SEGY format. Several attributes were applied to the image log data in order to complete such a process.
Attributes run on image logs include iso-frequency component, RMS amplitude (iterative), variance (edge method) (Fig. 9), graphic equalizer, generalized spectral decomposition, remove bias, sweetness, phase shift (Fig. 10), amplitude contrast, structural smoothing, first derivative, relative acoustic impedance (Fig. 11), trace AGC, trace AGC (iterative), trace gradient, reflection intensity, time gain (Fig. 12), gradient magnitude, local structural azimuth, t* attenuation, second derivative, local structural dip and chaos (Fig. 13).
The attributes shown in Figs. 9, 10, 11, 12 and 13 collectively enhance fracture interpretability, with notable variation in their detection capabilities. The iso-frequency component (Fig. 9) outperforms all others by identifying 45 fractures and offering superior clarity of fine-scale discontinuities. RMS amplitude (39 fractures) and variance (32 fractures) also demonstrate reliable performance. In contrast, attributes in Figs. 10 and 11, such as remove bias (38 fractures), first derivative (40 fractures), and relative acoustic impedance (36 fractures), provide moderate improvement in fracture visibility, though less refined than the iso-frequency output. Meanwhile, Figs. 12 and 13 include low-performing attributes like reflection intensity (8 fractures) and gradient magnitude (10 fractures), which show poor fracture enhancement and limited applicability. Overall, the systematic comparison across figures confirms that iso-frequency and a subset of signal-processing attributes significantly improve fracture clarity and detectability, supporting more accurate structural interpretation from SEGY image logs.
These attributes, particularly the iso-frequency component and edge-based variance, increased contrast between fractures and host rock matrix. Discontinuities, lineaments, and fracture traces features thus became more visually evident, so that more descriptive and confident interpretation could be achieved than from the unprocessed original image log.
Additionally, the iso-frequency component stands out as the most powerful in identifying fractures among the various attributes examined (Fig. 9). This feature has proven to be very effective at highlighting fracture patterns making it a useful tool for petrophysical and geophysical research. Its importance in interpreting image logs is highlighted by the fact that the iso-frequency component can distinguish fractures more clearly compared to the other attributes.
The iso-frequency component can detect rock properties or the presence of fractures. It is particularly effective at identifying subtle variations in seismic frequencies. This feature also lowers noise and improves the resolution of seismic data producing subsurface images of higher quality by concentrating on important frequencies and filtering out unwanted ones (Fig. 9).
The iterative RMS amplitude and the edge-based variance attributes demonstrated good performance in detecting fractures. Since fractures have distinct amplitude characteristics, these methods are particularly useful. Hence fractures frequently have higher amplitudes than the surrounding rock RMS Amplitude—which is computed as the average amplitude over a given window—highlights them. This is explained by the fact that a stronger signal results from seismic waves reflecting off the fracture surface. In contrast the variance (edge method) concentrates on the amplitude variation within a trace. Since seismic waves encounter distinct densities and properties at the fracture interface, they cause more noticeable amplitude fluctuations. Fractures therefore tend to exhibit more amplitude changes compared to the host rock (Fig. 9).
A) shows the SEGY image and the attributes that are run on the SEGY Image are: Iso-frequency component (B), RMS amplitude (iterative) (C), Variance (edge method) (D). Fractures interpreted from each attribute output are marked with dashed lines. Iso-frequency component detects 45 fractures, the highest among all attributes, offering exceptional resolution of fine-scale discontinuities. RMS amplitude identifies 39 fractures, emphasizing zones of elevated signal energy typical of fracture surfaces. Variance (edge method) reveals 32 fractures, delineating abrupt changes in amplitude associated with fracture edges.
Moderate-performing spectral and signal attributes with fracture traces marked: (A) Graphic equalizer – 26 fractures; (B) Generalized spectral decomposition – 26 fractures; (C) Remove bias – 38 fractures; (D) Sweetness – 30 fractures; (E) Phase shift – 24 fractures. Remove bias performs well by suppressing artifacts and enhancing relevant signals. Sweetness moderately enhances energy contrasts, while phase shift and spectral decomposition are less effective in fine-scale fracture isolation. Overall, this group shows moderate to fair detectability.
Several attributes showed moderate to weak performance in detecting fractures on SEGY images. These include amplitude contrast, structural smoothing, first derivative, relative acoustic impedance, trace AGC, iterative trace AGC, trace gradient, reflection intensity, and time gain (Figs. 11 and 12). This is explained by their reduced sensitivity to the particular features of fractures. For example, amplitude contrast although useful in emphasizing regions with notable amplitude shifts may not be sensitive enough to detect minute fracture-related variations. Likewise, the smaller more localized characteristics of fractures may not be sufficiently highlighted by structural smoothing which aims to improve geological structures and reduce noise.
Moreover, features such as the trace gradient and first derivative emphasize amplitude variations over depth or time which might not be enough to indicate the presence of fractures. The amplitude is adjusted along the trace-by-trace AGC and its iterative version which may conceal amplitude variations related to fracture. Time gain which is intended to compensate for signal attenuation may mask fracture-related amplitude changes while reflection intensity which gauges the strength of seismic reflections may not be specific to fracture signatures. These characteristics have limited sensitivity to minute fracture-related amplitude variations. This makes them less useful than other attributes for properly detecting fractures, even though they can still reveal important information about seismic data.
Structural and derivative-based attributes with varying fracture detection performance: (A) Amplitude contrast – 27 fractures; (B) Structural smoothing – 34 fractures; (C) First derivative – 40 fractures; (D) Relative acoustic impedance – 36 fractures. First derivative shows relatively strong performance, capturing rapid amplitude changes associated with fractures. Structural smoothing and impedance methods perform moderately well, though less sensitive to finer fracture features. Amplitude contrast ranks lower in overall detectability.
Normalization and gain-based attributes with limited fracture enhancement: (A) Trace AGC – 32 fractures; (B) Trace AGC (iterative) – 29 fractures; (C) Trace gradient – 36 fractures; (D) Reflection intensity – 8 fractures; (E) Time gain – 5 fractures. AGC methods moderately enhance amplitude consistency but tend to suppress subtle variations linked to fractures. Trace gradient performs better in highlighting depth-related amplitude trends. Reflection intensity and time gain show poor detectability, identifying the fewest fractures.
Because of these inherent limitations the attributes gradient magnitude, local structural azimuth, t* attenuation, second derivative, local structural dip and chaos typically perform poorly in detecting fractures on SEGY images (Fig. 13). These characteristics lack the sensitivity to detect small amplitude shifts and textural variations linked to fractures. Instead, they focus more on large-scale geological features or general signal complexity. For example, since azimuth and dip are better suited for larger-scale structural analysis gradient magnitude is more effective at highlighting sharp edges and boundaries. Analogously although t* Attenuation is sensitive to variations in the characteristics of the rock it is also subject to several influences which makes it challenging to separate signals unique to a given fracture.
Moreover, the second derivative may not appropriately capture the particular features of fractures even with its sensitivity to small variations. Since a multitude of other factors can also contribute to chaotic signals, the chaos attribute often captures a mix of effects. Although it is intended to detect complex spots in seismic data, it is not specific enough to reliably identify fracture signatures. These features are therefore typically less useful than other methods created especially to identify minute changes in amplitude and texture, which are necessary attributes of fractures on SEGY images.
In seismic data, reflector displacement is a main criterion of fractures however in image logs where fractures are frequently unrelated to layer shifts this feature is less noticeable. Attributes like chaos which mainly depend on layer shifts and discontinuities find it difficult to identify fractures in SEGY images due to this lack of displacement. As such attributes such as chaos may not be the best at detecting fractures in SEGY images because they are primarily concerned with detect shifts in reflectors which are less visible in image logs.
Low-performing structural and signal complexity attributes with weak fracture detectability: (A) Gradient magnitude – 10 fractures; (B) Local structural azimuth – 19 fractures; (C) t* attenuation – 29 fractures; (D) Second derivative – 17 fractures; (E) Local structural dip – 28 fractures; (F) Chaos – 16 fractures. These attributes show poor detectability overall, mainly capturing large-scale structural trends or signal irregularities rather than fine-scale fractures. Their low fracture counts reinforce their limited applicability for detailed fracture analysis on SEGY image logs.
The results of applying the iso-frequency component attribute to the SEGY image data displayed in Fig. 9 (A and B) are shown in Fig. 14. The modeling software requires setting the desired frequency and number of cycles parameters. These values should be optimized to properly highlight data fractures when using the iso-frequency component attribute. In seismic data the desired frequency identifies the precise frequency range that is most susceptible to fractures. In this instance, a value of 65 suggests that the most useful frequency range for identifying fracture-related signals is this one.
Additionally, the length of the wavelet used in the iso-frequency component analysis is defined by the number of cycles parameter which has a direct effect on the analysis resolution. A wavelet with a value of 1.5 cycles is thought to be almost short allowing for increased sensitivity to minute changes in the signal and a more precise and detailed representation of fractures.
The SEGY image (A) and the iso-frequency component attribute applied to the SEGY image log (B). This attribute effectively emphasizes fractures in the data by presenting them with a unique signature, enabling easy distinction from the surrounding rock.
To better demonstrate the impact of the iso-frequency component attribute in fracture identification, all three sets of images—legacy acoustic, and its corresponding SEGY and iso-frequency component—are shown using the same color display (Fig. 15). The legacy and SEGY images are very visually comparable, and this verifies the success of the SEGY conversion. In contrast, the iso-frequency component aspect significantly enhances the definition of features related to fractures, and structural detail is more apparent than in the original and SEGY images.
Comparison of legacy acoustic image (A), corresponding SEGY reconstruction (B), and iso-frequency component (C), all displayed using a unified color scale. While the legacy and SEGY images exhibit close visual agreement, the iso-frequency component enhances the visibility of fracture-related features and improves structural detail clarity.
In an attempt to better elucidate the ability of various seismic attributes to detect natural fractures, a detailed comparison table (Table 2) was constructed that listed the number of fractures and qualitative detectability rating for each attribute. Table 2 complements Table 1 by offering a more direct comparison of attribute performance based on visual interpretation and fracture traceability in the SEGY-transformed image logs. Among the 23 attributes analyzed, the iso-frequency component detected the highest number of fractures (45) and was rated as having fair detectability, standing out as the most effective in highlighting fine-scale discontinuities. Other attributes such as RMS amplitude (iterative), first derivative, and remove bias also demonstrated relatively good performance, with fracture counts above 35 and moderate detectability.
On the other hand, several attributes—including time gain, reflection intensity, gradient magnitude, and chaos—showed restricted capability in fracture identification, detecting fewer than 10–20 fractures and being categorized as poor in detectability. These results reinforce the observation that attributes tailored to enhancing frequency content (e.g., iso-frequency) or amplitude variation (e.g., RMS amplitude, edge-based variance) tend to be more sensitive to fracture features.
The abundance of fractures in the formation is highlighted in the petrophysical software display of the CBIL image from a different interval of well-A (Fig. 16). This implies that the region has experienced notable tectonic activity which has resulted in the formation of multiple fractures within the reservoir. These fractures may significantly affect the rocks permeability which may have an effect on how fluids move through the reservoir.
The fractures seen in Fig. 16 follow a consistent pattern with comparable angles of strike and dip. This implies that they are part of the same fracture set pointing to a common geological genesis and possibly affecting the properties of fluid flow.
The CBIL image from a different interval of well-A, displayed in the petrophysical software, reveals a high concentration of fractures within the formation. The consistent pattern of fractures with similar dip and strike angles suggests they are part of the same fracture set.
An increased number of fractures can be seen in the iso-frequency component attribute when comparing SEGY images with it. Furthermore, the iso-frequency component attribute (Fig. 17) provides a clearer definition of the fracture details.
The iso-frequency component attribute can provide more precise information about fracture networks than the SEGY image as demonstrated by the convincing comparison between the two shown in Fig. 17. There appears to be a dark patch in the SEGY image at the point where the red arrow points but it is not clear how many fractures there are. On the other hand, the iso-frequency component attribute reveals four separate fractures in this region suggesting more intricate fractures than what is first seen (Fig. 17).
The SEGY image points to a single fracture at the point indicated by the black arrow. A wider aperture in one of the fractures than the other is highlighted by the iso-frequency component attribute. Analogously the iso-frequency component attribute unmistakably reveals two distinct fractures at the location indicated by the blue arrow whereas the SEGY image only displays one fracture (Fig. 17). This consistent pattern indicates that the iso-frequency component attribute is better at identifying minute fracture features leading to a more thorough comprehension of the underlying geological structures.
The SEGY image (A) and the iso-frequency component attribute result (B). Comparison between SEGY image and iso-frequency component attribute highlights finer details of fracture networks, revealing multiple fractures where SEGY image suggests single fractures.
Additionally, FMI image logs served as a stand-in for electrical image logs when running seismic attributes (Fig. 18). For this reason, detailed images of the borehole wall including features like bedding planes, vugs and fractures are provided by the FMI logs that are obtained using a micro-resistivity imager. FMI logs use micro-resistivity measurements, which provide an alternative viewpoint on the features of the borehole wall. In contrast, electrical image logs rely on the electrical characteristics of the formation.
Well-B likewise shows extensive natural fracture development just like well-A. There are induced fractures in addition to natural fractures (Fig. 18). Fractures known as induced fractures are brought about by the stress that hydraulic fracturing or drilling operations cause. In the Asmari carbonate formation where natural fractures are primarily responsible for production these image logs of wells A and B are found.
There is a noticeable set of naturally occurring fractures in Fig. 18 that have a steep dip angle. These fractures show a consistent northeastern dip direction indicating that a localized stress field played a role in their formation. The high dip angle suggests that these fractures are inclined steeply which may lead to the formation of a network of connected pathways that improve fluid flow and reservoir permeability.
Comparing the iso-frequency component attribute to the standard FMI image it offers a more nuanced and detailed view of the natural fractures. This improved visualization shows more fractures at specific depth intervals with finer details that the standard image might have overlooked. Five fractures can be seen in the iso-frequency component image (Fig. 19) in a portion of the image log where two natural fractures are indicated by black arrows.
In addition, the iso-frequency component displays induced fractures or fractures that were generated by the drilling process (Fig. 19). Understanding the possible risks and advantages of hydraulic fracturing and other reservoir stimulation techniques requires knowledge of this information. The direction of horizontal stresses in the reservoir can be identified by using induced fractures as indicators. The direction of the principal stress that affected the formation of these fractures is usually indicated by the way they are oriented.
Fractures in well-B within the Asmari carbonate formation, characterized by a high dip angle towards the northeast, suggest they formed under the influence of a regional stress field.
The FMI SEGY image (A) and the FMI iso-frequency component attribute (B). While two natural fractures are indicated by black arrows on the SEGY image, the iso-frequency component image reveals five fractures. Induced fractures are shown with blue arrows on both images.
For better illustration of the effectiveness of the iso-frequency component in identifying fractures, all three image types, namely legacy electrical, its SEGY equivalent reconstruction, and the iso-frequency component, are displayed with the same coloring (Fig. 20). The high degree of similarity of legacy and SEGY images guarantee correctness of SEGY conversion. In comparison, the iso-frequency component provides enhanced clarity of fracture-related features, enhancing structural characteristics over both the original and the SEGY images.
Comparison of legacy electrical image (A), corresponding SEGY reconstruction (B), and iso-frequency component (C), all displayed using a unified color scale. The legacy and SEGY images show strong visual similarity, validating the SEGY conversion process. The iso-frequency component (C) further enhances the expression of fracture-related features and improves the clarity of structural details.
Conclusion
The substantial potential of combining seismic attribute analysis with SEGY image logs for improved fracture characterization is illustrated by this study. Conventional image log data were effectively integrated with advanced seismic processing methods through the conversion of electrical and ultrasonic logs into SEGY format.
Compared to other seismic attributes, the results demonstrate how well the iso-frequency component attribute can distinguish natural fractures by offering a more distinct representation. This characteristic allows geoscientists to gain a thorough understanding of the fracture networks that affect fluid flow and reservoir properties. It is especially effective at detecting minute variations in seismic frequencies related to fracture signatures.
The study emphasizes the importance of this integrated method in image logs processing procedure by applying seismic attributes to both naturally occurring and induced fractures. As a result, geoscientists can improve hydrocarbon recovery by managing reservoirs more effectively by selecting fracture networks based on their orientation, density and connectivity.
By supporting the use of SEGY image logs as a common tool for fracture analysis in the sector this research advances geophysical techniques. Future research should continue to explore the potential of such an integrated approach. This will help improve our understanding of subsurface features and support the discovery of new geological events, especially across various geological settings.
It is worth noting that the method detailed in this study is qualitative and mainly reliant on visual interpretation of attribute-enhanced image logs. While seismic attributes have been encouraging in detecting fracture-related events, attribute performance analysis was subjective qualitative comparisons rather than quantitative measures. As a result, interpretations may vary depending on the interpreter’s experience or visualization parameters. Future research should attempt to incorporate objective verification techniques to support and augment the qualitative findings presented here.
Data availability
All data applied to accomplish this research will be available under reasonable demand to the first author.
Code availability
Python code was written for this research.
References
Chopra, S. & Marfurt, K. J. Seismic Attributes for Prospect Identification and Reservoir Characterization (Society of Exploration Geophysicists and European Association of Geoscientists and Engineers, 2007).
Subrahmanyam, D. & Rao, P. H. Seismic attributes-a review. In 7th International Conference and Exposition on Petroleum Geophysics, Hyderabad, India (2008).
Oumarou, S. et al. Seismic attributes in reservoir characterization: an overview. Arab. J. Geosci. 14, 1–15 (2021).
Yilmaz, Ö. Seismic Data Analysis: processing, inversion, and Interpretation of Seismic Data (Society of Exploration Geophysicists, 2001).
Wang, K. Y., XU, Q. Y., Zhang, G. F., Cheng, M. C. & LI, P. H. Summary of seismic attribute analysis. Progress Geophys. 28 (2), 815–823 (2013).
Ashraf, U. et al. Application of unconventional seismic attributes and unsupervised machine learning for the identification of fault and fracture network. Appl. Sci. 10(11), 3864 (2020).
Russell, S. D., Akbar, M., Vissapragada, B. & Walkden, G. M. Rock types and permeability prediction from dipmeter and image logs: Shuaiba reservoir (Aptian), Abu Dhabi. AAPG Bull. 86 (10), 1709–1732 (2002).
Barton, C. A. & Zoback, M. D. Discrimination of natural fractures from drilling-induced wellbore failures in wellbore image data-implications for reservoir permeability SPE res. Eval Eng. 5, 249–254 (2002).
Ameen, M. S. & Hailwood, E. A. A new technology for the characterization of microfractured reservoirs (test case: Unayzah reservoir, Wudayhi field, Saudi Arabia). AAPG Bull. 92 (1), 31–52 (2008).
Xu, C., Cronin, T. P., McGinness, T. E. & Steer, B. Middle Atokan sediment gravity flows in the red oak field, Arkoma Basin, oklahoma: A sedimentary analysis using electrical borehole images and wireline logs. AAPG Bull. 93 (1), 1–29 (2009).
Donselaar, M. E. & Schmidt, J. M. The application of borehole image logs to fluvial facies interpretation. In Dipmeter and Borehole Image Log Technology (eds Poppelreiter, M. et al.) 145–166 (AAPG Memoir 92, 2010).
Folkestad, A., Veselovsky, Z. & Roberts, P. Utilising borehole image logs to interpret delta to estuarine system: A case study of the subsurface lower jurassic cook formation in the Norwegian Northern North sea. Mar. Pet. Geol. 29, 255–275 (2012).
Nie, X., Zou, C., Pan, L., Huang, Z. & Liu, D. Fracture analysis and determination of in-situ stress direction from resistivity and acoustic image logs and core data in the Wenchuan Earthquake Fault Scientific Drilling Borehole-2 (50–1370 m). Tectonophysics 593, 161–171 (2013).
Ameen, M. S. Fracture and in-situ stress patterns and impact on performance in the Khuff structural prospects, Eastern offshore Saudi Arabia. Mar. Pet. Geol. 50 (50), 166–184 (2014).
Kosari, E., Ghareh-Cheloo, S., Kadkhodaie-Ilkhchi, A. & Bahroudi, A. Fracture characterization by fusion of geophysical and Geomechanical data: a case study from the Asmari reservoir, the central Zagros fold-thrust belt. J. Geophys. Eng. 12, 130–143 (2015).
Xu, J., Zhang, B., Qin, Y., Cao, G. & Zhang, H. Method for calculating the fracture porosity of tight-fracture reservoirs. Geophysics 81 (4), IM57–IM70 (2016).
Brekke, H., MacEachern, J. A., Roenitz, T. & Dashtgard, S. E. The use of microresistivity image logs for facies interpretations: an example in point-bar deposits of the McMurray Formation, Alberta, Canada. AAPG Bull. 101 (5), 655–682 (2017).
Ezati, M., Olfati, H., Khadem, M., Yazdanipour, M. R. & Mohammad, A. Petrophysical evaluation of a severely heterogeneous carbonate Reservoir; electrical image logs usage in porosity Estimation. Open. J. Geol. 9 (3), 193–200 (2019).
Davari, M. A., Ezati, M., Jafarizadeh, F. & Motamedi, M. Enhancing breakout identification in geomechanical modeling: using fullset logs with machine learning in carbonate reservoirs. Earth Sci. Inform. 18(1), 81 (2025).
Goodall, T. M., Møller, N. K. & Rønningsland, T. M. The integration of electrical image logs with core data for improved sedimentological interpretation. Geol. Soc. Lond. Spec. Publ. 136(1), 237–248 (1998).
Donselaar, M. E. & Schmidt, J. M. Integration of outcrop and borehole image logs for high-resolution facies interpretation: example from a fluvial fan in the Ebro basin. Spain Sedimentology. 52, 1021–1042. https://doi.org/10.1111/j.1365-3091.2005.00737.x (2005).
Xu, C. Interpreting shoreline sands using borehole images—A case study of the cretaceous ferron sands in Utah. AAPG Bull. 91, 1319–1338 (2007).
Wilson, T. H., Smith, V. & Brown, A. Developing a model discrete fracture network, drilling, and enhanced oil recovery strategy in an unconventional naturally fractured reservoir using integrated field, image log, and three-dimensional seismic data. AAPG Bull. 99 (4), 735–762 (2015).
Zoback, M. et al. Determination of stress orientation and magnitude in deep wells. Int. J. Rock Mech. Min. Sci. 40, 1049–1076 (2003).
Babcock, E. A. In-situ stress magnitude and orientation estimates for cretaceous coal-bearing strata beneath the plains area of central and Southern Alberta. Bull. Can. Pet. Geol. 51, 1–28 (2003).
Laubach, S. E., Olson, J. E. & Gale, J. F. W. Are open fractures necessarily aligned with maximum horizontal stress? Earth Planet. Sci. Lett. 222 (1), 191–195 (2004).
Prioul, R., Donald, A., Koepsell, R., Marzouki, Z. E. & Bratton T Forward modeling of fracture-induced Sonic anisotropy using a combination of borehole image and Sonic logs. Geophysics 72 (4), 135–147 (2007).
Rajabi, M., Tingay, M., Heidbach, O., Hillis, R. & Reynolds, S. The present-day stress field of Australia. Earth Sci. Rev. 168, 165–189 (2017).
Massiot, C., Mcnamara, D. D. & Lewis, B. Processing and analysis of high temperature geothermal acoustic borehole image logs in the Taupo volcanic zone. New. Z. Geothermics. 53, 190–201 (2015).
Ezati, M., Azizzadeh, M., Riahi, M. A., Fattahpour, V. & Honarmand, J. Wellbore stability analysis using integrated Geomechanical modeling: a case study from the Sarvak reservoir in one of the SW Iranian oil fields. Arab. J. Geosci. 13, 1–19 (2020).
Saadatnia, N., Sharghi, Y., Moghadasi, J. & Ezati, M. Coupled hydro-mechanical simulation in the carbonate reservoir of a giant oil field in Southwest Iran. J. Petroleum Explor. Prod. Technol. 14 (1), 59–83 (2024).
Laubach, S. E. Practical approaches to identifying sealed and open fractures. AAPG Bull. 87 (4), 561–579 (2003).
Zeng, L. B. & Li, X. Y. Fractures in sandstone reservoirs with ultra-low permeability: A case study of the upper triassic Yanchang formation in the Ordos Basin, China. AAPG Bull. 93 (4), 461–477 (2009).
Ameen, M. S., MacPherson, K., Al-Marhoon, M. I. & Rahim, Z. Diverse fracture properties and their impact on performance in conventional and tight-gas reservoirs, Saudi arabia: the Unayzah, South Haradh case study. AAPG Bull. 96 (3), 459–492 (2012).
Zeng, L., Su, H., Tang, X., Peng, Y. & Gong, L. Fractured tight sandstone oil and gas reservoirs: A new play type in the Dongpu depression, Bohai Bay Basin, China. AAPG Bull. 97 (3), 363–377 (2013).
Ezati, M., Soleimani, B. & Moazeni, M. S. Fracture and horizontal stress analysis of Dalan formation using FMI image log in one of Southwestern Iranian oil wells. J. Tethys. 2 (1), 1–8 (2014).
Nian, T., Wang, G. & Song, H. Open tensile fractures at depth in anticlines: a case study in the Tarim basin, NW China. Terra Nova. 29 (3), 183–190 (2017).
Ezati, M., Azizzadeh, M., Riahi, M. A., Fattahpour, V. & Honarmand, J. Characterization of micro-fractures in carbonate Sarvak reservoir, using petrophysical and geological data, SW Iran. J. Petrol. Sci. Eng. 170, 675–695 (2018).
Zeng, L. et al. Natural fractures and their contribution to tight gas conglomerate reservoirs: a case study in the northwestern Sichuan Basin China. J. Pet. Sci. Eng. 210, 110028 (2022).
Zhao, T., Che, X., Qiao, W. & Yang, B. Field application of near-borehole fracture acoustic imaging using a three-dimensional scanning inversion method. Geoenergy Sci. Eng. 240, 213070 (2024).
Wood, D. A. Expanding role of borehole image logs in reservoir fracture and heterogeneity characterization: A review. Adv. Geo-Energy Res. 12(3) (2024).
Ameen, M. S., Buhidma, I. M. & Rahim, Z. The function of fractures and in-situ stresses in the Khuff reservoir performance, onshore fields, Saudi Arabia. AAPG Bull. 94 (1), 27–60 (2010).
Wilson, M. E. J. et al. Development of a Papua new Guinean onshore carbonate reservoir: a comparative borehole image (BHI) and petrographic evaluation. Mar. Pet. Geol. 44, 164–195 (2013).
Keeton, G., Pranter, M., Cole, R. D. & Gustason E.R Stratigraphic architecture of fluvial deposits from borehole images, spectral-gamma-ray response, and outcrop analogs, piceance Basin, Colorado. AAPG Bull. 99 (10), 1929–1956 (2015).
Lai, J. et al. A review on the applications of image logs in structural analysis and sedimentary characterization. Mar. Pet. Geol. 95, 139–166 (2018).
Bordenave, M. L. & Hegre, J. A. The influence of tectonics on the entrapment of oil in the Dezful Embayment, Zagros Foldbelt, Iran. J. Pet. Geol. 28 (4), 339–368 (2005).
Bordenave, M. L. & Hegre, J. A. Current distribution of oil and gas fields in the Zagros Fold Belt of Iran and contiguous offshore as the result of the petroleum systems. Geol. Soc. Lond. Spec. Publ. 330(1), 291–353 (2010).
Stocklin, J. Possible ancient continental margins in Iran. In The Geology of Continental Margins 873–887 (Springer Berlin Heidelberg, 1974).
Berberian, M. & King, G. C. P. Towards a paleogeography and tectonic evolution of Iran. Can. J. Earth Sci. 18 (2), 210–265 (1981).
Salati, S., van Ruitenbeek, F. J., Carranza, E. J. M., van der Meer, F. D. & Tangestani, M. H. Conceptual modeling of onshore hydrocarbon seep occurrence in the Dezful Embayment, SW Iran. Mar. Pet. Geol. 43, 102–120 (2013).
Shaban, A., Sherkati, S. & Miri, S. A. Comparison between curvature and 3D strain analysis methods for fracture predicting in the Gachsaran oil field (Iran). Geol. Mag. 148 (5–6), 868–878 (2011).
Rezaie, A. H. & Nogole-Sadat, M. A. Fracture modeling in Asmari reservoir of Rag-e Sefid oil-field by using multiwell image log (FMS/FMI). Iran. Int. J. Sci. 5 (1), 107–121 (2004).
Darvishzadeh, A. Geology of Iran: stratigraphy, tectonic, metamorphism, and Magmatism (Amir Kabir Publication Institute, 2009).
Abbasi, A., Kodkhodaie, A., Mahari, R. & Moussavi-Harami, R. Maximum entropy spectral analysis of gamma ray logs for cyclostratigraphic analysis of the late Albian-Early turonian Sarvak formation in the Anaran exploration block, Southwestern Zagros, Iran. Earth Sci. Inf. 17 (1), 1–20 (2024).
Daws, C., Olivero, A., Tripakis, S. & Yovine, S. Borehole Geology, geomechanics and 3D reservoir modeling brochure. Oilfield Rev. 1, 1–12 (2002).
Funding
This research is supported by the research grant of the University of Tabriz (grant number 4830), which we are grateful of.
Author information
Authors and Affiliations
Contributions
Mohsen Ezati: Writing & edition, original draft preparation, Software, Methodology, Computation.Ali Kadkhodaie: Conceptualization, Methodology, Review & Edition, Software.Vahid Bolandi: Conceptualization, Methodology, Review & Edition, Software.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Ezati, M., Kadkhodaie, A. & Bolandi, V. Exploring the potential of extracting seismic attributes from image logs for enhanced fracture characterization. Sci Rep 15, 38044 (2025). https://doi.org/10.1038/s41598-025-21949-9
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1038/s41598-025-21949-9
