3D structural modelling and spectral decomposition analysis of Jay Field offshore Niger Delta, Nigeria

Abstract

In the Niger Delta, the Jay Field offshore structural model and seismic attribute analysis were carried out to enhance prospect evaluation and mitigate risk in field development operations, especially in regions with few wells and poor seismic data. Integration of a 3D seismic cube and one well with composite logs was used to carry out this analysis. Horizon and fault mapping was enhanced with Seismic attribute analysis, and an optimum frequency for spectral decomposition was identified using spectral analysis. The 3D structural framework revealed collapsed crest structures and rollover anticlines which act as hydrocarbon traping systems on the field. Detailed spectral analysis was carried out before using the spectral decomposition attribute to identify the most suitable (spectral) frequency at which the possible reservoir zones become more pronounced. The spectral frequency of 25Hz has been analysed to give high amplitude in the hydrocarbon accumulated region (indicated around the well position). This was used to evaluate prospective regions. The spectral decomposition attribute analysis predicts regions of possible quality reservoirs. This paper reveals that seismic attributes such as spectral decomposition, energy, and similarity can improve reflection patterns in poor to good quality seismic data. In this paper, the energy and spectral decomposition are veritable attributes that have improved reflection patterns and had allowed to identify anomalous amplitude associated with hydrocarbon accumulation in this paper.

Introduction

Seismic attributes analysis effectively predict reservoir heterogeneity and uncertainties when building models and identifying prospective regions. Enhancing and predicting lithology, structures, and fluid properties of the reservoir by integrating wells and seismic attribute analysis is employed in the oil and gas sectors to reduce drilling risk [1, 2]. The spatial distribution of Structural trends, stratigraphic events, reservoir properties, and hydrocarbon prospects can be resolved by integrating seismic amplitudes with frequency variation attributes [2, 3].

It is vital to enhance the continuity of seismic reflectors using attribute analysis when interpreting seismic stratigraphy and depositional regimes [2, 4]. In Multi-attribute analysis, the distribution of various depositional environments, such as river channels, can be reliably predicted [2, 5]. However, attributes should be correlated to detect if different attributes will not enhance the same result because using too many attributes can generate geostatistical errors in hydrocarbon reservoir properties prediction [2, 4]. An attribute should not be used because it is available as a software tool. The ultimate step, especially for areas where well log data is unavailable, is an accurate and effective estimation of facies distribution and prediction of hydrocarbon reservoirs properties based on seismic patterns recognition and correlation [6].

Several authors have applied seismic attributes to identify reflector continuity, direct hydrocarbon indicators (DHIs), enhance faults and interpret potential hydrocarbon reservoirs [1, 3, 7]. Moreover, poor quality 3D seismic data can achieve a good result by integrating Seismic attributes such as spectral decomposition and variance attributes in channel identification and facies modelling of hydrocarbon reservoirs [5].

In summary, Seismic interpretation using attribute analysis are visualisation technique for plotting amplitudes and ameliorating the signal-to-noise ratio of seismic data [7]. This study verifies that the structural model and seismic attribute analysis on Jay Field offshore Niger Delta are substantive tools to enhance prospect evaluation and reduce risk in field development operations in regions where sparse well is available.

Geological settings

The Niger Basin Basin within the Gulf of Guinea (Fig. 1), a continental margin in equatorial West Africa, covers a 75km² area with a clastic fill of about 12km and borders the Atlantic Ocean between latitudes 5° 52′ 50″ and 6° 15′ 00″ and longitude 4° 81′ 25″ and 4° 92′ 25″ [9]. It forms one of the world’s central hydrocarbon provinces, with proven ultimate recoverable reserves, of 34.5 billion barrels of oil and recoverable reserves of 93.8 trillion cubic feet of gas [8, 10]. The Tripple junction led to the evolution of the Cenozoic Niger Delta Basin, developed in the Late Jurassic (Fig. 2) when South America and African plates separated [12]. Within the Niger-Delta Basin, gravity tectonics is the leading cause of deformation after the rifting ceases; this internal deformation is caused by the mobility of the shales [8]. The dominant structural styles in the Niger-Delta Basin are Growth faults caused by deeply buried, over-pressured mobile shale and slope instability [8]. The deltaic deposit within the basin is prograding southwestward; at active regions, it forms depobelts at each development stage from the Eocene to recent (Fig. 1) [8]. Shale diapers, rollover anticlines, collapsed crest, clay-filled channels, and synthetic and antithetic faults are the prominent trapping structures within the Niger-Delta Basin (Fig. 3), and these structures are expressions of the gravity tectonics in each depobelt within the Niger-Delta Basin [8, 14, 15]. The building of the structural framework in the basin is by growth faults that also bound some depobelts, and the hydrocarbon traps within the basin are mostly rollover structures associated with these growth faults, although sometimes dissected by crestal faults [16]. The Niger Delta stratigraphy section is categorized based on sand-shale ratios and divided into three Formations: The Akata Formation, which underlies the entire delta, is the potential source rock and comprises thick shale sequences formed when terrestrial organic matter and clays, deposited during low stand in deep water environment (Fig. 4) [13, 17]. The primary hydrocarbon-bearing reservoir unit is the Agbada Formation, which overlies the Akata Formation; it consists of paralic siliciclastic sediment deposited in the delta-topset, delta front, and fluvio deltaic environment. The third Formation is the Benin Formation which overlies the Agbada Formation; it consists of recent alluvial and upper coastal plain sand with a thickness is about 2000 m [18].

Fig. 1

A Map of the Niger Delta. B Various depobelts and locations of the study area (Jay Field) [8]

Fig. 2

A Location of Nigeria. B Early cretaceous separation of Africa and South America showing the proposed Niger Delta Triple Junction. C Map showing Mesozoic spreading rates (cm/year) for Africa-South American separation [11]

Fig 3

Stratigraphic sequence model for the central portion of the Niger Delta showing the relation of source rock, migration pathways, and hydrocarbon traps related to growth faults [13]

Fig. 4

Stratigraphic column shows the three formations of the Niger Delta [8]

Methodology

The data available for the study are (i) 3D seismic data, (ii) one well (Fig. 5) with suites of logs (Caliper, Gamma-ray, Resistivity, Neutron, Density, Sonic, Vshale, and Sw), and (iii) check shot data. This dataset was provided by Chevron Nigeria through the Nigerian Department of Petroleum Resources (DPR) specifically for research and academic purpose.

Fig. 5

Identified lithologies and hydrocarbon contact on well Jay 1 log

Well log interpretation

The gamma-ray log was used for lithology identification because high gamma readings indicate shale in response to its high radioactive mineral contents, and low gamma-ray indicates Sand lithology. The hydrocarbon-bearing sand unit lies within the Agbada Formation, composed of sand and shale intercalation sequences. The gamma-ray and resistivity log identified four hydrocarbon-bearing sand units, sand 1 to 4 (Fig. 5), and the lowest known contact is the oil-water contact.

Porosity estimation

The four pay zones’ porosity was calculated using the density logs (Eq. 1). where Pma is the bulk density of the sand matrix (2.65g/cm³), Pb is the measured bulk density log, and Pf is the density of the fluid [19].

$$\Phi =\frac{P_{ma}-{P}_b}{P_{ma}-{P}_f}$$

(1)

Net to gross estimation

The thickness of shale streaks within the reservoir interval was deducted from the total reservoir column. Therefore, the hydrocarbon-bearing thickness ratio to the total reservoir thickness is net to gross [20].

Water saturation estimation

Water saturation was estimated using the Simandoux model (Eq. 2), which considers the shale volume of the sand reservoir [21]. Where R_t is the true formation resistivity or resistivity of the formation, R_w is the resistivity of water, S_w is the water saturation, V_sh is the volume of shale in the rock, F is the formation factor and R_sh is the resistivity for shale.

$$\frac{1}{R_t}=\frac{S_w^2}{F.{R}_w}+\frac{V_{sh}{S}_w}{R_{sh}}.$$

(2)

Seismic interpretation

The seismic data is a 3D pre-stack migrated Segy data covering an area of o about 16.3 by 9.7km² offshore Niger Delta (Fig. 6). The seismic interpretation covers seismic-to-well calibration, fault interpretation, hydrocarbon horizon definition, seismic attribute analysis, and structural modelling.

Fig. 6

The 3D seismic survey layout and Jay-1 well

Seismic to well tie

The sonic and density logs generated a synthetic seismogram to tie the well to the seismic. The sonic and density logs first generate the Acoustic Impedance, and a reflection coefficient was computed using the Zoepritz equation [22]. It was convolved with extracted wavelet from the seismic (zero-phase wavelet) to generate the synthetic seismogram compared with the seismic data correlated with the most significant events. The identified pay zones correspond to high amplitude events or peaks on the seismic data. The check shot data equation was used as the time-to-depth relationship for time and depth maps (Fig. 7).

Fig. 7

Plot of check shot data used as the time-depth relationship

Seismic attribute analysis

Seismic attributes aid in mapping faults and horizons [23]. Attribute analysis was carried out on a time slice to reveal geomorphological features such as fault planes and trends [24]. Various attributes were analysed, and some have proved functional while others lack consistency. Attributes employed using the OpenDTect software are spectral decomposition, similarity, and energy attributes.

Spectral decomposition

Spectral decomposition is a method which uses a correlation between a user-defined cosine frequency and autocorrelation function of the original seismic data. It may indicate subtle lithologic features which are not detected in original amplitude and identify prospect regions of hydrocarbon accumulation. It can also be used to interpret depositional elements.

Similarity

Similarity measures how consistent is the orientation of a reflector estimated based on local structural azimuth attribute. It enhances edges, salt bodies and chaotic patterns within the original seismic data. It detects edges, such as faults and discontinuities.

Energy

Energy is the modulus, or the total instantaneous energy, of the complex seismic trace. It distinguishes stratigraphic, lithologic, and fluid lateral variations inside a hydrocarbon reservoir and major changes of lithology and sequence boundaries. It can also be used as a DHI, it may identify bright spots.

Spectral decomposition analysis

The Jay Field’s frequency spectrum ranges from 1 to about 125Hz (Fig. 8), of which the dominant frequency band is within 5 to 60Hz. Spectral decomposition was carried out to identify the (dominant) frequency at which the hydrocarbon region (observations for the water region were equally carried out) is mainly represented by bright amplitude on the seismic data. The seismic data was decomposed at frequencies of 5-Hz intervals and the amplitude of the hydrocarbon region in sand 1 to 4 was measured at each of the frequency intervals. The amplitude was plotted against the selected frequency (Fig. 17) to identify the specific frequency at which the amplitude of the hydrocarbon region will be enhanced on the seismic data.

Fig. 8

Spectral decomposition attribute frequency spectrum of Jay Field

Fault and horizon mapping

The seismic data’s faults and identified pay zone horizons were mapped on every tenth inline and crossline to generate time maps. The time maps were converted to depth maps with the time-depth relationship function.

Data integration

The well log and seismic interpretation are integrated to estimate the hydrocarbon pore volume and rank prospects [25]. The volumetric analysis was limited to hydrocarbon pore volume to avoid ambiguous and uncertain reserve estimation due to impassable values of the formation volume factor (FVF) and recoverable factor (RF) from analogue data. The equation utilised in estimating the hydrocarbon pore volume is.

$$\mathrm{HCPV}=k\times H\times A\times \phi \times \left(1-{S}_w\right)\times \mathrm{NTG}$$

(3)

where k is the unit conversion factor 6.28 for units in meters, H = reservoir thickness, A = area of prospect, Φ = porosity, and NTG = net to gross.

Results and discussion

Seismic attribute analysis

Various attributes have been applied, but a few have shown convincing results. The attributes applied have made mapping of hydrocarbon horizons and faults easier and more accurate despite the poor to fair quality of available seismic data. The results of the functional attributes in the region are presented (Figs. 9 and 10).

Fig. 9

This image shows how the fault was enhanced using the spectral decomposition attribute, and the Similarity attribute made the fault more visible. a Default seismic. b Spectral decomposition (15 to 40 Hz). c Similarity

Fig. 10

Enhanced horizons using the energy attribute. a Default seismic. b Energy attribute

The similarity volume attribute has proven helpful in fault mapping because it makes faults more pronounced around the pay zones where the seismic resolution is poor (Fig. 9). Good fault mapping is invaluable in defining the geometry and architecture of traps in the reservoir model.

The energy volume attribute has proved helpful in the stratigraphic interpretation. Attenuation has strongly diminished the amplitude of the default seismic (Fig. 10) and has resulted in events smearing. However, energy attributes have made the horizon more visible and continuous for regions with difficulty in horizon mapping.

Spectral decomposition of the seismic data with the frequency band of 15 to 45Hz has proven to enhance the visualisation of the fault geometry (Fig. 9b).

Well log interpretation

The well log interpretations (Fig. 11) have identified four pay zones named sand 1 to 4. Cut off of 7Ωm and 0.5 have been applied to resistivity and water saturation, respectively (the third and sixth track). The shaded green region is the hydrocarbon-filled portion of the reservoir. The summarised result of parameters such as water saturation Sw, Porosity, and net to gross is presented in Table 1.

Fig. 11

The hydrocarbon-bearing reservoirs identified on well Jay 1

Table 1 Calculated petrophysical parameters for reservoirs in well Jay 1

Seismic interpretation result

The faults and top of the hydrocarbon-bearing reservoirs identified in the wells were mapped across the seismic cube (Fig. 12). The mapped horizons in time were depth converted to depth maps by the time-depth relationship presented in Figs. 13, 14, 15, and 16. It can be observed that the contour patterns and structures on the time map are also represented on the depth map, which further indicates that the velocity model used is applicable.

Fig. 12

Seismic interpretation of fault and hydrocarbon horizons in Jay Field

Fig. 13

Time and depth map of sand 1

Fig. 14

Time and depth map of sand 2

Fig. 15

Time and depth map of sand 3

Fig. 16

Time and depth map of sand 3

Spectral decomposition analysis

Spectral decomposition was carried out on various time slices within the hydrocarbon-bearing interval around well Jay 1 at various frequencies (5 to 75Hz) with a 5Hz interval to identify the frequency (associated dominant frequency) at which the hydrocarbon-bearing region (as identified by the well) will give the highest amplitude. The amplitude value around the well was plotted against (various) frequencies (Fig. 17). The highest amplitude observed across the spectral plot of the time slices cutting across the hydrocarbon column is about 25Hz (Fig. 17). The time slice is spectrally decomposed at 25Hz, and the high amplitude event (similar to that around the well) can be associated with possible hydrocarbon accumulation. Around the well, it can be observed that a high amplitude event is dispersed in the default seismic display; however, its geometry is more defined in the spectral decomposition attribute display when a frequency of 25Hz was applied to delineate the possible hydrocarbon zones (Figs. 18 and 19). Likewise, it can be observed that the dominant frequency for the water-filled reservoir region is about 30Hz (Fig. 15). The high amplitude anomaly indicates in other regions the possible hydrocarbon accumulation.

Fig. 17

Plot of amplitude against frequency at different frequency spectra on each time slice

Fig. 18

Default seismic and spectral decomposition attributes of sand 1 and sand 2

Fig. 19

Default seismic and spectral decomposition attributes of sand 3 and sand 4

Structural model

The depth model of the reservoirs mapped on the seismic cube is displayed in Figs. 20 and 21. The depth model shows the varying thickness of the reservoir sands. The structural framework of the reservoir was developed by depth converting the faults and mapped surfaces with the well as a constraint. The reservoir depth model serves as a primary input for further modelling. The depth reservoir model can be upscaled with various petrophysical parameters from wells. However, well Jay 1 alone may not be fit to populate the model. In the future, this model can be upscaled when more wells are drilled on identified prospects (Fig. 20). Figure 21 shows the sand 1 reservoir, fault depth model, and collapsed crest structures with adjoining antithetic and growth fault (also displayed in the seismic section).

Fig. 20

a Cross-section of the mapped reservoir depth/thickness model. b 3-D view cross-section of mapped reservoir thickness model

Fig. 21

Structural model of collapsed crest structure on sand 1

Volume estimation and prospect ranking

In the identified prospects, fault-assisted traps are considered low risk because the integrity of the trap does not depend on the sealing or non-sealing property of the rock. Fault-dependent traps are considered a medium risk because the trapped accumulation of hydrocarbon may be compromised if the fault is non-sealing. Other prospects that may serve as a trap for accumulated hydrocarbon provided closure exists outside the survey area. They are considered to have a high risk because these traps’ integrity is unknown. The unrisked reserves were risked based on the geologic chance of success, and prospects were ranked based on structures and attribute support (Table 2). Prospects that are fault assisted, fault dependent, and potential traps provided that closure existing outside the study area is identified as a low, medium, or high-risk prospect (Fig. 22). Prospects with high anomalies similar to the well region have been ranked higher than those with the same structures [26, 27]. Most prospects are present on the four sand reservoirs and may be penetrated by a well (similar to Jay 1); therefore, prospects ranking in (Table 2) are sums of the reserves within the reservoirs. The displayed reserves are based on a P50 estimate. The average estimated hydrocarbon pore volume (HCPV) ranges from 19 to 342MBLs, and the total HCPV estimate is 634MBls (Table 2). Prospect 7 has the lowest risk (ranked 1), while prospect 6 has the highest risk (ranked 9) (Fig. 22).

Table 2 Estimated reserves and ranking in Jay Field

Fig. 22

Juxtaposed attribute and depth map generalised for Jay Field to integrate attribute-supported structures in prospect ranking

Conclusions

Structural modelling and seismic attribute analysis of Jay Field, in the Niger Delta offshore of Nigeria, has been carried out to enhance prospect evaluation and risk assessment. The available data are a 3D seismic cube, one well with a suite of logs, check shot, and data acquisition map. However, the 3D seismic cube reflection is of poor to fair quality. Seismic attribute analysis improved the available poor to fair quality seismic reflection data while mapping the faults and horizons. Seismic attribute analyses have also been proven to improve reflection patterns and identify anomalies that may be associated with hydrocarbon accumulation. Various attributes have been employed during this study, but spectral decomposition, similarity, and energy have good results. Similarity attributes have helped in mapping fault. Spectral decomposition and energy attributes are veritable attributes in improving reflection patterns and anomalous amplitude associated with hydrocarbon accumulation. Spectral decomposition analysis has shown that the data spectrum ranges from 1 to 125Hz. The spectral frequency of 25Hz has been analysed to give high amplitude in the hydrocarbon accumulated region (indicated around the well position) and utilised in evaluating prospective regions.

Identified pay zones in the well are labelled sand 1 to sand 4. The hydrocarbon-bearing reservoirs have an average porosity range of 21 to 26%, average hydrocarbon saturation range of 66 to 72%, thickness range of 24 to 114m, and net to a gross range of 0.54 to 0.91. The traps mapped from the seismic data interpretation are structural with fault-assisted and dependent closures. About ten prospects have been identified in the study area. Some prospective regions have been suggested from their contour pattern, provided closure exists outside the survey area. The average estimated hydrocarbon reserves in the Jay Field range from 19 to 342MBLs. The total reserve estimate is 634MBLs with an estimated standard error range of ±12% to ±28%. Seismic attribute analyses have proven invaluable in mapping both fault and horizon, which is a prerequisite to the architecture of the reservoir structural model. Spectral decomposition analysis has proved instrumental in inferring prospective regions.

Recommendation

Seismic attributes such as spectral decomposition, energy, and similarity may improve reflection patterns and fault detection in poor to good seismic data offshore Niger Delta. Detailed spectral analysis should be carried out before using the spectral decomposition attribute to know the (spectral) frequency that will make the reservoir more pronounced.