Introduction

The pursuit for the accumulation of hydrocarbons required an advance comprehensive approach to expedite exploration, appraisal, development and production activities (Evans et al. 2019). However, the capability to sufficiently characterize reservoirs remains a critical challenge. It is vital to use dependable reservoir characterization approaches for an effective and reliable reservoir delineation and description. The reservoir zonation approach is one of the methods used by researchers to understand the internal architecture and factors controlling flow zones in a reservoir (Nabawy and Al-Azazi 2015; Nabawy and ElHariri 2008; Opuwari et al. 2019; M. Opuwari et al. 2020a, b; Radwan et al. 2020).The reservoir zonation approach helps to understand reservoir heterogeneity and identify productive zones within a reservoir, which is subsequently upscale to field-scale reservoir mapping (Opuwari 2010; Qassamipour et al. 2021).

Several case studies using core, seismic, and wireline log data have been effectively applied in reservoir characterization to explore and develop hydrocarbon(Das and Chatterjee 2018; Gogoi and Chatterjee 2019; Gupta et al. 2012; Nabawy et al. 2018; Neely et al. 2021; Sohail et al. 2020; Teama et al. 2019; Verwer and Braaksma 2009). An enhanced classification of lithofacies and depositional environment centred on core, well log, and seismic data is important for formation evaluation, prospect identification and field development (Ashraf et al. 2019; Elamri and Opuwari 2016; Evans et al. 2019). The heterogeneity and the complex nature of reservoir properties can be effectively described and quantified by means of available core data, well log and seismic data (Gogoi and Chatterjee 2019).

The study area is the Y field (Fig. 1), which has three exploration wells (Y-01, Y-02, Y03) situated in the central Bredasdorp Basin offshore of South Africa (Brown 1995; McMillan et al. 1997). An additional exploration well(X-01) in Field X, approximately 6 km from the Y field, is used for broad correlation. The primary target of this study is in the Lower Cretaceous (Valaginian to Albian age) sediments where hydrocarbon has been discovered within the 5At1 to 13At1 interval (Petroleum Agency 2003).

Fig. 1
figure 1

Location map of study are showing wells used in this study in red circle

Full size image

Extensive work has been done on the geology and petroleum system of the Bredasdorp Basin that is bounded by the southern Outeniqua basin, of which recent hydrocarbon was discovered. In one of those studies in the central Bredasdorp Basin, it was documented that some gas reservoirs are thin, which are cumbersome or unidentifiable on a seismic scale (Corbett et al. 2011; Grobbler 2005; Turner et al. 2000). Previous studies in the Bredasdorp Basin on reservoir characterization include (Baiyegunhi et al. 2020a, 2020b, 2020c; Barton 1997; Mudaly et al. 2009; Opuwari et al. 2020a, b; Opuwari et al. 2021; Saffou et al. 2020). Although there is an in-house study by an oil and gas company on the reservoir architecture description and characterization of the Bredasdorp Basin, there is still a gap in the published literature on comprehensive reservoir characterization of the central part of Bredasdorp Basin. Previous seismic data acquired in the study area have poor resolutions, making it challenging to create plausible geological models.

This study is to provide solution to this knowledge gap by using the newly acquired seismic data with an enhanced resolution, integrating core and well logs. The study has produced a calculated volume of gas in the 10At1 and 13At1 sand and identified areas to focus during further field development programme because of the good reservoir quality. Improved seismic resolution enhanced the accuracy of results. Consequently, this paper presents importance of an integrated approach to reservoir characterization.

Geological setting

The offshore basins of South Africa are grouped into three sections: the western section comprises the Orange Basin, the Eastern area (Zululand and Durban Basins), and the Southern region (Outeniqua Basin), as shown in Fig. 1. The Outeniqua Basin, which is the focus of this study, was formed earlier than the west and east Gondwana separation and stretched on the Falkland plateau (Jungslager 1999; Roux 2000). There are four sub-basins in the Outeniqua basin known as the Bredasdorp (study area), Gamtoos, Pletmos and Algoa. According to Miall (1980), the Southern offshore basins exhibit rift basins with normal faults and half grabens.

Rifting in Southern Africa is postulated to begin during the middle to the late Jurassic period (Boada et al. 2001; McMillan et al. 1997). The rifting that occurred created normal faults (Roux 1997). The sediments that have mostly filled the Bredasdorp Basin are marine sediments deposited in the Late Jurassic to Early Cretaceous fluvial and shallow marine deposits of the Aptian to the Maastrichtian age (Grobbler 2005; Turner et al. 2000), presented in Fig. 2.

Fig. 2
figure 2

Chronostratigraphic chart of the Bredasdorp Basin

Full size image

During the mid-Jurassic to the Valanginian, synrift I was manifested from the basement to the 1At1 unconformity (Fig. 2). Rifting initially was due to horst and graben structures characteristic of extensional tectonics (Elliott 1997; McMillan et al. 1997). In the Late Valanginian to the Hauterivian, 1–6At1 sequence boundary is observed, which depict the synrift II phase, characterized by swift subsidence and extensive flooding (Van der Spuy 2003). The 6–13At1 sequence boundary is observed in the transitional early drift phase at Hauterivian to the Aptian. In the early Aptian, there was a decline in subsidence rates and faulting, resulting in a more stable Bredasdorp Basin (Petroleum Agency 2003). The drift phase occurred at the Albian to Maastrichtian time which is the 13–22At1.

Most oil and gas source rocks are shales in the central Bredasdorp Basin, positioned among the transitional period in deep marine, transgressive, and early highstand tracts of the synrift 1 (Fig. 2). The sandstone reservoirs in the central Bredasdorp Basin are predominantly upper shallow marine (USM) depositional environment. The upper shallow marine sandstones are the best gas reservoirs and have good porosities and permeability (Broad et al. 2006).

Data and methods

The Petroleum oil and gas company of South Africa (PetroSA) and the Petroleum Agency of South Africa (PASA) provided the data set used in this study. The data set provided includes suites of conventional wireline logs (gamma-ray, calliper, density, sonic, neutron, and resistivity), seismic data (2D and 3D lines), core data, conventional and special core analysis reports, and well completion report. Schlumberger Petrel and Sedlog software packages were used for data analysis and interpretation.

Two of the wells used (Y-01 and Y-02) had core data laid out and watered to observe the grain size and structural features. This process was accompanied by a recording of lithology description of the core to establish sedimentology parameters such as grain size distribution, structure and lithofacies distribution. The cores were photographed, and the recorded data from the core evaluation were exported to Sedlog software. A project was created in Petrel, and the various data sets obtained were loaded in their different formats before quality check and control were performed. Quality control and data editing were done to improve data quality where necessary and used for analysis and interpretation.

For petrophysical evaluation and stratigraphy modelling, a wireline log was integrated with core and seismic data. Stratigraphic modelling allows the determination of similar rock bodies at different locations where well data are available. The petrophysical evaluation process commences with the estimation of reservoir parameters, followed by determining the volume of clay, total and effective porosity, water saturation, and permeability estimation using core facies as a guide.

The volume of clay (Vcl) was determined from the GR log by calculating the gamma-ray index using the equation developed by Altas (1979), and Fertl and Frost (1980), as follows:

$${\text{GRI}} = \, \frac{{{\text{Vcl}} = {\text{Gr }}\left( {{\text{log}}} \right) \, {-}{\text{ Gr }}({\text{min}})}}{{{\text{Gr }}\left( {{\text{max}}} \right) \, {-}{\text{ Gr }}\left( {{\text{min}}} \right)}}$$
(1)

where Gr (log) = Gamma-ray log value of formation to be evaluated, Gr(min) = Minimum value of gamma-ray reading on a clean formation, Gr (max) = Maximum gamma ray reading. For a clean sand interval, calculated Vcl ≤ 10%; for a shaly interval, Vcl is between 10 and 35%; and for a shaly interval, Vcl ≥ 35% (Opuwari 2010).

Porosity was determined from the density log using the following relationships of Asquith and Gibson (1982) as follows;

$$\begin{aligned} {\text{Porosity }} = & \, \rho {\text{max}} - \, \rho {\text{b}} \\ \rho {\text{max}} - \, \rho {\text{f}} \\ \end{aligned}$$
(2)

where ρmax = Matrix density determined from core, ρb = Bulk density, ρf = Fluid density.

The effective porosity was determined by correcting the effect of shale on porosity using the following equation:

$${\text{Effective porosity }} = \frac{{\left( {\rho {\text{max}} - \, \rho {\text{b}}} \right) \, {-}{\text{Vcl }}\left( {\rho {\text{max}} - \, \rho {\text{cl}}} \right)}}{{\left( {\rho {\text{max}} - \, \rho {\text{f}}} \right) \, {-} \, \left( {\rho {\text{max}} - \, \rho {\text{f}}} \right)}}$$
(3)

where ρcl = Density of clay.

The Indonesia water saturation (SWIndonesia) equation proposed by Poupon-Leveaux (1971) calculated water saturation. The equation is provided below:

The empirical relationship can be written as follows:

$${\text{Swind}} = \left( {{1}/\sqrt {Rt} = \frac{{\sqrt {\Phi {\text{e}}^{{\text{m}}} } }}{{\text{a*Rw}}} + \frac{{{\text{Vcl}}^{{\left( {{1} - {\text{Vcl}}/{2}} \right) }} }}{{\sqrt {Rcl} }} \times {\text{ Sw}}^{{{\text{n}}/{2}}} } \right)$$
(4)

where Rt = Resistivity curve from log, Rcl = Resistivity of wet clay, Φe = Effective porosity calculated from Eq. 3. Swind = Water saturation, fraction, Vcl = Volume of shale, fraction. Rw = Formation water resistivity. m = Cementation exponent. a = Tortuosity factor. n = Saturation exponent.

A relationship obtained from the cross-plot of core permeability against porosity was used to determine the permeability from the log as follows:

$${\text{Permeability}} = {1}.{146 }*\Phi {\text{d}}^{{0.{67} }} \left( {{\text{R}}^{{2}} = 0.{93}} \right)$$
(5)

where Φd = Porosity from density log.

The seismic interpretation process commenced with seismic to wireline log tie, Mistie adjustment, then unconformities identification, faults picking and horizon mapping.

Various surfaces and models were created in the workflow that culminates into the volumetric calculation. The reservoirs were delineated using the properties in a 3D grid, and the gas water contact (GWC) was also established. The flowchart of the study is presented in Fig. 3.

Fig. 3
figure 3

Flow chart of study

Full size image

Results and discussion

Core analysis

The result of the core logging revealed four distinct lithofacies (Fig. 4) classified as follows:

Fig. 4
figure 4

Result of core logging showing four lithofacies

Full size image

Facies 1: Greyish black, fine-grained claystone, well sorted and laminated.

Facies 2: Intercalation of claystone and fine sandstone, fair sorting, claystone laminated.

Facies 3: Grey, massive sandstone, medium-grained, fair sorting, coarsening upward sequence.

Facies 4: Grey, fine to medium grain, fair sorting, sandstone interbedded with clay, coarsening upward sequence.

The four classified lithofacies are present in all the studied wells. Facies 1, predominantly of claystone, has the lowest rock quality and dominates in the upper part of the wells between 14 and 13A, intercalated with fine sandstone at intervals (Fig. 4). Facies 4 and 3 have the best rock quality with a reasonable thickness in all the well, bounded at the top and bottom by facies 1.

Well correlation

The lithostratigraphy correlation of the geological horizons based on the gamma-ray log is presented in Fig. 5. A baseline of 75 American Petroleum Institute (API) units was used to differentiate between a sandstone interval (0-75API) and a shale interval (75-150API). The top (at depth 2849 m) and bottom (at depth 2856 m) of horizon 13AT1 sandstones were correlated. The 13At1 sandstone has similar thicknesses throughout the three wells in the field.

Fig. 5
figure 5

Correlation of geological horizons based on gamma ray log

Full size image

Below the 13At1 sandstone is the 10At1 sandstone (from about 2898 m as top, to about 2955 m as base). Three key reservoir sandstone units were identified between the top (marker 10At1) and bottom (marker 10At1) of 10At1 sandstone in the studied well. The 10At1 horizon represents the base of the 10At1 sandstone. Well Y-01 has similar sandstone bodies to well Y-03. The sandstone bodies are thicker in well Y-01 than Y-02, and lower gamma-ray values are displayed in these sandstone bodies.

Facies modelling was performed from the gamma-ray log using the following criteria:

Facies 1: 75–150 API.

Facies 2: 25-75API.

Facies 3: 0–15 API.

Facies 4: 15-25API.

Due to the heterogeneity observed in the sediments, a layering operation was performed to incorporate the very fine sandstones found between the claystone and the intercalation of facies. The layering method resamples the facies data into 3D cells, takes the dominant facies in the pre-selected interval, and assigns it to the cell penetrated by the well (Puskarczyk 2020). The results of the facies layering method are compared to the core identified facies of the wells which are presented in track 5 of the wells (Fig. 6). The upscaled facies obtained from the layering method matches the core facies. Facies are important in reservoir modelling because the petrophysical properties are often highly correlated with facies types. Knowledge of facies constrains the range of variability in reservoir properties such as porosity, permeability, and saturation (Elamri and Opuwari 2016; Magoba and Opuwari 2017).

Fig. 6
figure 6

Results of facies layering method compared to core identified facies

Full size image

Seismic interpretation

An essential step in integrating seismic data measured in the time domain with that of well log data measured in depth domain is the seismic well tie operation, where well logs, especially the density and sonic logs, are used to calibrate seismic data, which has a lower vertical resolution than the well logs (Bader et al. 2019). The density and sonic logs are used to generate a time-depth relationship involving synthetic seismic trace combined to create an acoustic impedance (White and Simm 2003).

An example of a seismic well tie for well Y-01 showing the seismic log trace is shown in Fig. 7. It is observed in Fig. 7 that a poor to medium quality tie between the seismic traces and log exists on the seismic section, which is the best we could achieve. The trough, denoted by blue colour, and the peak by red colour match the seismic reflection section very well. The trough showed a point of zero crossing from negative to positive. The top-down approach was adopted for mapping the horizons for velocity modelling from the seabed down to the 10At1 horizon. The results of the horizon mapping showed two horizons of interest of good reservoir sandstone overlaid with claystone at the top and bottom, which are the 13At1 and 10At1 horizons. The 13At1 sandstone presented variation in thickness with high acoustic impedance values that creates a peak wavelet (Corbett et al. 2011), as illustrated in Fig. 8. The 10At1 horizon was easier to map because it was more visible than the top sand. An arbitrary line was drawn from well Y-02 –Y01- Y-03 (Fig. 8).

Fig. 7
figure 7

Example of a seismic well tie for well Y-01 showing seismic log traces

Full size image
Fig. 8
figure 8

Random lines show seismic section with mapped horizons from 22 to 10At1

Full size image

The faults do not have huge displacements and therefore do not show much of shift in the horizon. Consequently, fault polygons were not used as input because there are no visible faults intersecting the surfaces with significant displacement. Surfaces are created in two-way time (TWT) units (Figs. 9a–c). The constructed surfaces in time were converted to depth using an average velocity model. The created depth surfaces are presented in Fig. 10a–b, while the velocity surfaces for horizon 13AT1 and 10At1, which are the horizons of interest, are displayed in Fig. 11, respectively. The studied wells are enclosed by 2860 m contour, and an evident increase in the depth towards the NE and SE direction is observed.

Fig. 9
figure 9

Results of created two-way time (TWT) surfaces

Full size image
Fig. 10
figure 10

Results of depth models for 13At1 and 10At1

Full size image
Fig. 11
figure 11

Results of velocity model for horizon 13At1 and 10At1

Full size image

The 3D structural model was created using the structural gridding method for property modelling suitable for the population of petrophysical models. The structural grid results (Fig. 12a) show two zones, which are zone 1(between 14 and 13At1 surfaces representing the 13At1 sandstone), and zone 2 (between the 13 and 10At1 surfaces that represents the 10At1 sandstone). Zone 2 appears to be much thicker than zone 1. Further investigation of the two zones by the layering method produces a detailed surface geometry, showing a fill structure (Fig. 12b). The results of the petrophysical models populated in a created structural grid are presented in the next section.

Fig. 12
figure 12

Structural grid results

Full size image

Volume calculation and uncertainty analysis

Continuous variables with spatial correlation, such as porosity, were populated stochastically using sequential Gaussian simulation (SGS) algorithm conditioned to the facies model using Petrel Software. Data analysis and transformation were performed as it is the main prerequisite to run the petrophysical modelling. Firstly, log-scale porosity was scaled-up, and quality checked into the pre-defined layers.

Effective porosity curve derived from petrophysical interpretation was used as input data for the porosity model. Statistical analysis of the porosity, permeability, and water saturation log data was performed, upscaled, and populated in the grid created earlier to produce a property model. Results of the upscaled values are presented in Table 1, while the distribution of porosity and facies is shown in Fig. 13a, b, and permeability and facies model is shown in Fig. 14a, b. The properties models are done conditioned to facies, to have a match between property values and facies. Comparison of property model with facies showed that facies 1, predominantly claystone, have low porosity and permeability values in Figs. 13 and 14. Conversely, sandstone exhibits higher porosity and permeability values. The porosity data of the field of study range between 5 and 12%. It is observed that facies 4 has the highest porosity values. The permeability surface model result (Fig. 14b) shows that the upper sequence (13At1 sand and above) of the formation has lower permeability values (0.5–10mD) in comparison to the lower sequence (10At1 sand) with permeability values ranging from 0.5 to 90mD. These high permeability values exhibited at 10At1 sand indicate well-connected pore spaces that will enhance reservoir performance (Opuwari et al. 2021, Goigo and Chatterjee 2019; Mondal and Chatterjee 2021). The water saturation models (Fig. 15a, b) with facies were executed separately for the 13At1 sandstone and Fig. 15c, f for 10At1 sandstone because they have different gas water contacts (GWCs). Figure 15a, which reveal fair to high water saturation between 40 and 80%. High water saturation is observed in the western side of the field of study, and a decline is observed towards the east.

Table 1 Results of the upscale petrophysical properties
Full size table
Fig. 13
figure 13

a, b Distribution of porosity property and facies

Full size image
Fig. 14
figure 14

a, b Results of the distribution of permeability with facies

Full size image
Fig. 15
figure 15figure 15figure 15

a, b Results of water saturation model for 13At1. c, d Results of water saturation model for the upper 10At1sand. e, f Results of water saturation model for the lower 10At1 sand

Full size image

Within the 10At1 sandstone, low to fair water saturation (40–65%) is revealed (Fig. 15c, e) for the upper and lower 10AT1 sand. The 10At1 sandstone shows water saturation ranging between 40 and 60%, which indicates the high deliverable of hydrocarbon in this sandstone. Similarly, the 10At1 sandstone model shows that the water saturation is higher at the western side of the field and lower at the eastern side of the field. A comparison of water saturation values for the different sequences showed that the claystone has higher water saturation than the sandstone, presented in Table 2. Generally, examining the other petrophysical models for the different sequences indicates that the 10At1 sequence has good pore volume and connectivity in the reservoir, and high deliverability of gas is expected.

Table 2 A comparison of water saturation values for the different sequences
Full size table

The surfaces created, indicating the gas water contacts (GWCs) displayed in red colour (Fig. 16a–c) was subsequently used to estimate the gas initially in place. The pink colour in Fig. 16a indicates the position of GWC used to determine the spill points. The GWCs were quality checked and validated by integrating core data, well logs and well completion reports. The GWC for the 13At1 sequence was identified at depth 2580 m, for 10At1 upper sand at depth 2889 m, and that of the lower 10At1 sand at depth 2926 m, used for calculating the initial gas in place.

Fig. 16
figure 16

a-c Position of gas water contact (GWC) used to determine the spill points a for 13 AT1 sand b upper 10At1 sand and c lower 10At1 sand

Full size image

Another parameter used to estimate the volume of hydrocarbon in place is the bulk rock volume (BRV) that determines the volume and the area of rock above the GWC. In calculating the volume of hydrocarbon (gas) initially in place, the bulk rock volume, net to gross (NTG) ratio, porosity, gas formation volume factor (Bg), and gas saturation of each zones were considered. The volume of gas in place is calculated from Eq. 5 as used by Ma (2018) and Aly (1989):

$$\begin{aligned} &{\text{Volume of gas in place }}\left( {{\text{GIIP}}} \right) \\&\quad = \frac{{{\text{BRV }}*{\text{ NTG }}*{\text{ porosity }}* \, \left( {{1} - {\text{Sw}}} \right) \, *{\text{ constant}}}}{{{\text{Gas Formation Volume Factor }}\left( {{\text{Bg}}} \right)}} \end{aligned}$$
(6)

The calculation of the volume of gas in place for the various zones (13At1,10At1 upper,10At1 lower) sand intervals is presented in Table 3. The 13At1 zone has the most volume of gas, 27.22 bcf, followed by the 10At1 lower with 17.48 bcf, and the 10At1 upper zone with 12.53bcf, a cumulative 57.23 bcf of gas calculated for all the zones.

Table 3 Results of the calculation of the volume of gas in place for the various zones
Full size table

The petrophysical parameters (porosity, water saturation and NTG) used to calculate the gas volume in place showed lower uncertainty and more reliability than the Petrel software's results using the core analysis method. For instance, an error in the measurements of the input parameters may result in inaccurate determination of the volume of hydrocarbon. The calibration of log measurements using core measurements reduced uncertainty and enhances the model's reliability by changing the petrophysical input parameters' values that help determine the best parameters that honour the model. The study has demonstrated how 10% error in the petrophysical calculations does not significantly change the initial volume in place results. The results confirm how a threshold within 10% error would not significantly affect the volume calculation results. However, errors higher than 10% in the input parameters would substantially influence the results of volumetric. The bulk or gross rock volume estimated from seismic is indicated as the parameter that may impact volumetric results.

Comparison of the volume in place between PetroSA and current study

The current study is compared with the unpublished in-house study done in 2005 by PetroSA before the seismic was reprocessed. The comparison between the two results is presented in Table 4. The volume in place of the current study differs from the study done in 2005, as different seismic data were used to generate depth maps and ultimately create models used to calculate volumetrics. In the 2005 PetroSA study, the 10At1 sands were divided into three channels. Only the upper and lower sections of the 10At1 sands were identified in the current study. Therefore, both the 13At1 sands and 10At1 sands combined have a higher volume in place in the current study compared to the study conducted in 2005.

Table 4 Comparison between a study done by PetroSA in 2005 and the current study
Full size table

The total volume in place has increased by almost twice the previous study. Therefore, it can be concluded that the seismic resolution has improved as mapping of events has become much visible, which increased the volume in place.

Conclusion

The primary purpose of this study was achieved, which is the estimation of the volume of gas in place of the 13At1 and 10At1 sandstones reservoirs deposited in the upper shallow marine environment. Results from this study revealed four facies grouped as facies 1 which is claystone, facies 2 is fine sand, facies 3 is fine-medium sand, and facies 4 is medium sandstone. The medium sandstone has the best reservoir quality, while the claystone has the lowest reservoir quality. This study has produced a calculated volume of gas in the 10At1 sand (upper and lower is 30.01bcf) higher than that of the 13At1 sand, 27.22 bcf of gas. A total volume of gas in place of 57 bcf was calculated for all the zones evaluated. The depth maps show high structural areas at the 13At1 and 10At1 events. It is advisable to focus further field development in the eastern part of the field because of good reservoir quality and production developed due to the sand in the deeper 10At1 sequence.