1

DIFFERENTS APPROACHES IN FRONTIER AND MATURE BASINS CASE STUDIES IN ALGERIA SOUTH WESTERN SAHARA PLATFORM

Madjid Badsi*, Lounes Adour**, Tayeb Tenkhi**, Marie Christine Cacas***

* Main author (Sonatrach Exploration Division) ** Co authors (Sonatrach Exploration Division) *** co author (Institut Français du Pétrole)

Keywords: Algeria, facies changes, fractured zones, workflow

1 Introduction In Algeria south-western basins, situated in the north west of the Hoggar-Shield (fig 1), gas fields are the main discoveries. Even the proven gas reserves are important, they are trivial compared to those generated and expulsed which were evidenced by geochemical studies.

To be able to prove more and more reserves, geologists have to work in both frontier and semi mature basins.

In order to investigate some hypotheses, we used different approaches to understand the deformation process especially with examples in frontier basins and modelling in the more mature basins. These latter are more concerned by predicting the facies changes and the fractured zones, building the discrete model of fractures network, constructing the workflow from seismic to simulation of the fluid flow and finally estimating the reserves.

These approaches combine several methods, including acoustic impedance, facies determination, statistics rules, often related with fractal behaviour of fault families, ante tracking analysis of fractures, and workflow from seismic to simulation.

2 Objectives of the paper

In this paper are described the main problematic posed in south western part of Algerian Sahara Platform in both frontier and mature basins, the approaches used to solve problems suggested by interpreters, and the main results which are considered quite satisfactory depending on the quantity of data and their quality.

In the frontier basin, the main problems are related to the lack of models or to their relative uncertainties. The case studies in such frontier basins are related to complex geologic area such as syn sedimentary faults or tilted blocks.

In the more explored basins, an important volume is proven with important gas fields, where a large commercial production was recorded from the Ordovician and Devonian reservoirs. The main problematic is related to the complexity of the tectonics, the lateral variations of facies, the inconsistency of production and the misunderstanding of the spatial distribution of fractures.

The paper shows also the contribution of the new acquisition techniques and seismic processing to assess models in the frontier basin. When studying mature basins however, the acoustic impedance is essential in modelling porosities; the fractal analyses and ante tracking help to 3D model fracture networks. This work stresses that combined approaches are necessary for reservoir development and for an increase in an upside gas potential in both frontier and mature basins; these latter needing further appropriate technologies.

2

3 Development methods

a) Stratigraphic Setting During the Palaeozoic, from the Cambrian to the Namurian, the sedimentation was mainly silicoclastic, related to various depositional environments including continental, marine, glacial and coastal deposits. These deposits lie upon the Infra-Tassillian unconformity (Fig 2). The source rocks in the basin are essentially the basal Silurian and Frasnian/Givetian hot shales. Some secondary source-rock potential may occur in the median part of the Silurian and Cambro-Ordovician sections. The main reservoir rocks consist of the Devonian and Cambro-Ordovician sandstone, which present poor primary petrophysical characteristics. These properties are, however, improved by natural fracturing. The sequence stratigraphic concepts applied for the late Ordovician cross section, carried out from cores analysis and out crop field studies, showed up two sedimentary cycles, transgression / regression (fig.3) of advance and retreat of the ice cap. These cycles, equivalent to the El Golea and M`Kratta units, are composed respectively by trangressive and regressive sequences of the 2

rd Order.

The global tendency of the late Ordovician section is transgressive, related to the global melt of the glacial late Ordovician inlandsis, located at the South of the studied area.

b) Structural evolution

The present structural style (fig4a) has resulted from its location, on the former Pan-African (750-550 Ma) orogenic craton and from the principal tectonic phases the basin has undergone.

The structuring of the basin began at the end of the upper Devonian followed by a strong subsidence (fig4c), but the main structures have been formed consequently to the major orogeny hercynian phase (fig 4b).

Based on previous studies, we use the (fig5) to summarise the different fractures observed, their orientation and their relationship with the constraints.

The Hercynian phase would have began with an E-W to NW–SE compression, which resulted in the regional N70-N110 joints, observed on both outcrops and wells (fig 6). The compression was continuing and created the NW-SE reverse faults, folds axis as well as joints. The frequency of the joints increases with the bending of the folds; they are termed “the joint of extrados”.

At last, one late folding phase could be at the origin of the N40 systematic joints but this is not marked everywhere. Therefore, it is the Hercynian phase that reshaped the basin and yet, later movements, probably during the Cretaceous, are noticeable on the seismic sections.

c) Presentation of Case studies

Example 1(frontier basin)

The Reggane basin is situated to the south of the Ougarta ridge (fig.1). It is affected by several tectonic events and subsidence which led to the deposition of more than 6000 m of sediments in the axial part of the basin. Dolerite sills lie upon the upper Paleozoic series. The seismic profile shown in figure 7 partly crosses the basin and may exhibit fictive structural features in the central part of the basin. The use of the rays paths model highlights the existence of a velocity anomaly and the geological

structures seen in the seismic profile do not actually exist. The corrected velocity model shows a

concordance in the Palaeozoic layers.

3

Example 2 (frontier basin)

The Bechar basin is located in the north eastern side of the Ougarta ridge. Figure 9 shows a Tournaisian depositional anomaly within the anticline structure in the western side of the seismic section. We can notice a differentiation in the structural style between the carboniferous and the ante-carboniferous sequences in one hand, and an unusual tremendous thickness of the carboniferous section in the other hand. Is it due to eustatism or tectonic extension during the Tournaisian, creating tilted blocs in a roll over fashion? These questions are still without any answer but a recently acquired 2D seismic provided a clue to the solution as it showed thin skinned tectonics and over thrusting that affect the carboniferous section mainly. We can observe in figure 10 these features within the carboniferous witch overlie unconformably the folded and tilted blocs underneath.

Structure 1 (mature basins)

It is a large anticline structure located in the central Ahnet basin (fig1. The structural map (fig.11) carried out at the top of the Ordovician target shows three main axes: -NW-SE faults marked by an average vertical displacement of 100 m and folded axes of a same orientation. -N70°-N80° reverse faults. - N40° fold axes. The wells located along the N-S and NE-SW axes produce gas whereas the only well in the western part is dry. A question arisen from this observation is whether the productivity is related to facies changes or to fracturing considerations.

Structure 2 (mature basins)

It’s a large structure also located in the central Ahnet basin (fig1). It’s a large elongated anticline bordered by N-S – oriented faults on its eastern flank and a series of faults on the southern and north-western flanks. The structure area is about 600 km

2, and assuming a water-oil contact, 200 km

2 may

contain saturated hydrocarbons.

The anticline is characterised by two distinctive compartments (fig.12): A foot-wall compartment down –lifted with an average throw of 700m, A hanging- wall compartment up-lifted in the west showing some local curvature axes and affected by a more restricted number of faults.

According to the obtained results the drilled boreholes situated at the vicinity of the N-S oriented fracture band are dry whereas those drilled far away from these faults have shown some poor to good oil flows. This statement leads to ask two main questions: Why the fractures do not enhance the oil productivity in the eastern compartment? Why the flows are random in the western compartment?

d) Presentation of Methods

Sand box modelling

This supplementary tool is used with the aim of understanding the deformation. It’s an experimental technique based on a rheological model of lithosphere with brittle (sands) and ductile (silicone) behaviour fig 13 By simulating a deformation on the basis of the geological knowledge and by imposing the good conditions on the limits, the solution is taken into account provided that it is not the only one because depending on the rheological parameters.

4

Numerical modelling

The numeric model was built from a current geologic cross section based on the surface geology, the seismic sections and the well data. Numerical analyses were made including a linear elastic behaviour and a non linear behaviour of sedimentary rocks. The effect of a listric fault which is generally related to the formation of the structure, on the spatial and temporal distribution of the fracturing rocks was studied.

Geometrical modelling. The analysis was made on the structure 1 (fig 11). Two types of geometrical modelling had been studied: - The probabilistic modelling of the faults (fractal analysis), the philosophy being to investigate a fractal object by being interested in more and more fine details. The probabilistic modelling takes into account all the parameters of the faults in the seismic scale as well as the fractal dimension to generate faults which are not solved by the seismic (small scale). - The analysis of the curvature of reservoirs horizons is used to localise zones with strong curvature and to determine the orientations of the associated fractures. Furthermore, we have the possibility of estimating the occurrence probability of opening fractures according to their angle with the maximum horizontal stress.

Reservoir modeling

The geo-cellular grid (fig14) used in the geological modeling is characterized by the following parameters:

- 04 zone representing the number of individualized compartments of studied area. - Grid cells: 200×200. - Number of cells (nI×nJ×nK): 97×123×164, - Number of nodes (nI×nJ×nK): 98×124×165, - Make zone:

o 02 zones to represent sequences I and II of cycle I : Zones 1 and 2, o 01 zone to represent cycle II : Zone 3. The 3D porosities model was carried out by using stochastic method "Sequential Gaussian Simulation (SGS)" and deterministic method "co-kriging", integrating the acoustic impedance as a secondary property. The results of the variogrammes analysis were integrated in the 3D modeling porosities workflow. The distribution of porosities in 3D space of studied area, integrating acoustic impedance as secondary properties, is at least 50% deterministic. In fact, the correlation coefficient estimated between the acoustic impedance at wells studied and porosities measured from cores equal to 0.6.

Fractures modeling The 3D sub seismic faults modeling was carried out from the 3D seismic cube with using a specific workflow by generating a succession of seismic attribute: smoothing, coherence, Ant tracking and automatic extracted fault.

5

4 Results

Outcrops and well results.

Outcrop observations fig 6 allowed acquiring some parameters relative to the attributes of the fractures. These results have to be completed. This will allow acquiring geologic rules to better adjust the three-dimensional network of the fractures.

The technique of the well imaging allowed to observe the characteristics of fractures and to observe the open fractures privileged directions.

We observed the variation of the fracturing according to the lithology of the encountered layers and the structural dipping. It is noticed that the fracturing is generally very important in sandy layers and very weak in the shale. The more the structural dip is strong, the more the chance of encountering fractures is high. It is due to the trajectory of the well (vertical line). The individual distinction of every fracture plan allowed to classify fractures in three directional families and to consider that the density of fracturing is similar for every family of fractures. Three fractures families also observed in the outcrops are analysed and their parameters are considered for the calculation of the real spacing between fractures and for the 3D representation of the fractures network.

Sand box modelling results.

The various experiments show that it is possible to understand the formation process provided that appropriate model was adopted. Several experiments show a structure similar to that observed in the south-western central basins. The last one using the hypothesis of the transpression is the one which well simulated the structure (fig15). It seems that the structure is a flower structure initiated on an anticline ramp. The western zone of the fault appears more folded than the rest of the structure. The strike hypothesis (transpression) can also explain the faults throws difference between the surface and depth horizons. The role of Silurian shale represented by a silicone layer is also highlighted in all the experiments except in the models with weak thickness. The Silurian shale seem to be determining in the distribution of the initial thrusting and the associated retro thrusting

Numerical modelling results.

The analysis of the results shows that a single compression fig 16 a (without basement fault) is improbable. The results of the model are indeed in contradictions with the observations. On the other hand, the hypothesis of propagation or detachment fold seems more probable (fig 16b). This suggests that the secondary faults can have a post hercynian origin or may correspond to the reactived hercynian faults.

Geometric modelling results.

The various tests of geometrical modelling allowed to choose the most realistic network of sub seismic faults and to distinguish the zones where the spatial distribution of this type of faults is the more dense (fig 17a) . This analysis allowed also to localise zones with strong curvature and to determine the orientations of the associated fractures (fig 17b).

The calculation of the spacing between fractures also allowed to adjust the fractures network (fig 18a), and to determine the direction of the gas flow around the wells. This simulation showed the importance of fractures NW-SE and diaclases N100 and their contribution to the gas flow (fig 18c).

The models show also the zones where it is possible to predict the fractures related to the curvature of the folds and their orientation, by using Gaussian and directional curvature. It is, thus, possible to determine the fractures, which could be reopened by the present horizontal maximum stress.

6

Reservoir modelling results

The result sequence stratigraphy concepts applied to upper Ordovician section showed up two complete sedimentary cycles transgression / regression corresponding respectively to El Golea and M`Kratta units, bounded at their base by important incised unconformities. The main results studied in this integrated approach are related to the stochastic simulation of the reservoir properties and to the deterministic modeling of the sub-seismic faults and fractures from 3D seismic cube, using a specific workflow generated by the application of some seismic attributes. The acoustic impedance variation is directly related to the petrophysic variations of the rocks. Integrated approach applied in this study was included variogram data analysis, conceptual model deposits, sequence stratigraphy concepts, seismic attributes, petro physics formation evaluation and PVT data. The calculation of reserves estimated on basis of 3D geological parameters model takes account the heterogeneities of the reservoir.

Fractures modeling results The deterministic modeling of the sub seismic faults applied on 3D seismic cube showed up a coherent fractures network with those studied on outcrops and on imagery of some wells. The analysis of the fractures modeling results showed up that the high concentration fractures at different scale are related to zones of maximum curvature. The analysis of distribution of these sub seismic faults throughout the main study field shows a closer relationship between fractures density, the curvature axes of the folds and the productivity. All these results have an impact on the planning of the future delineation wells and in the exploration-production strategy.

7

Conclusion

This study allowed testing several methods to improve our knowledge in both frontier and mature basins.

The main problems in the frontier basin are mainly related to the lack of models or to their relative uncertainties. The case studies in such frontier basins are related to complex geologic area such as syn sedimentary faults or tilted blocks. Nevertheless, some approaches led to confirm certain hypothesis. In these frontier basins, the new acquisition techniques or processing such as ray tracing demonstrate the effectiveness of the assumptions.

In the mature basins, this study allowed testing several methods to improve our knowledge of the deformation process, the spatial distribution of the fractured rocks, and the stochastic simulation of the reservoir properties.

The sand box analyses have allowed to answer the questions related to the formation of these large anticlines.

The results of sand box modelling show that this type of structure is generated only, because of the pre-existing discontinuity in the basement. It appears, in fact, to have developed as ramp anticline from the pre-existing normal faults of the basement.

The numerical modelling shows that the fracturing associated with this type of structures is located in their Western part (when the pre-existing fault plane is supposed in the same compartment).

The results of geometric modelling of the fractures network indicate that the sub seismic faults are located near the major faults. The main results studied in the integrated approach (structure 2) are related to the stochastic simulation of the reservoir properties and to the deterministic modeling of the sub-seismic faults and fractures from 3D seismic cube, using a specific workflow generated by the application of some seismic attributes. The calculation of reserves estimated on basis of 3D geological parameters model takes account the heterogeneities of the reservoir. The deterministic modeling of the sub seismic faults applied on 3D seismic cube showed up a coherent fractures network with those studied on outcrops and on imagery of some wells

To improve the geologic knowledge in the mature basin, in particular the geometry of the fault network, the efforts should concentrate on the seismic acquisition and processing.

In zones with proved potential of hydrocarbons, 3D seismic acquisitions (3D azimuth) will bring more knowledge and better information of the fault network and the facies change. .

The use of D.S.I tools of type can inform about the intensity of the fracturing and the azimuth of the horizontal stress surrounding of the wells. Walk away can help to show more clearly the intensity of the fracturing around the wells. This enables to make extrapolations up to the reservoir scale. Outcrops observations are necessary to acquire geologic rules to simulate the most plausible fracture network. Acquisitions of well data by the imaging techniques as well as by oriented cores are also necessary and will help to better constraint the fractures network and to simulate the flow in the producing structures of the basin.

8

References Badsi, M, 1993. Failles et fractures dans le bassin de l’Ahnet. Apport du retraitement sismique. Mémoire D.E.A, Géosciences E.N.S.P.M , Université Pierre et Marie Curie. Badsi, M, 1998. Fracturation naturelle des roches : Application au bassin de l’Ahnet (ALGERIE), thèse de Doctorat de l’Université Pierre et Marie Curie Beekman, F, Badsi, M, In situ stress analysis in the Ahnet basin, Tectonophysics ; August, 1996 Beuf, S., Biju-Duval, B., De Charpal, D., Rognon, R. & Bennacef, A., 1971. Les grès du Paléozoïque inférieur au Sahara. Sédimentation et discontinuités : évolution structurale d’un craton. Publications institut Français du pétrole, Collège sciences et technologie du pétrole 18, Technip ed. Paris, pp 464 . Beghoul, M.S, 1991, apport et contribution de l’analyse des diagraphies à la connaissance d’un bassin sédimentaire. Application au bassin de Timimoun. Thèse de Doctorat de l’Université Louis Pasteur Strasbourg (France) Boudjemaa, A., 1987. Evolution structurale du bassin pétrolier triasique du Sahara nord oriental (Algérie). Unpublished PhD. Thesis, Université Paris Sud. Fabre J., 1979. Introduction à la géologie du Sahara. SNED. Fabre J., 2005. Géologie du Sahara occidental et central. Musée royal de l’Afrique centrale, Tervuren, 2005. Gringarten Alain C., 1986. Computer – Aided Well Test Analysis. SPE 1986 International Meeting on Petroleum Engineering held in Beijing, China March 17 -20, 1986. Gringarten Alain C., 1987. How To Recognize "Double - Porosity" Systems From Well Tests. SPE16437. Rahmani, A ; Semmad, A ; Aggoun, N. synthèse des résultats pétroliers de la structure de Bahar el Hammar. Octobre, 1996 (rapport interne) Robertson Group, 1989. Algerian petroleum geology and hydrocarbon potential. Sonatrach - Beicip Franlab, 2003. 3D fracture modeling and fracture properties computation on Ahnet Field. Rapport interne SONATRACH.

Sonatrach - Gustavson Ass., 2000. Ahnet – Timimoun basin study. Takherist.D, 1990 Structure crustale, subsidence Mesozoique et flux de chaleur dans les bassins Nord Saharien Algérie. apport de la gravimétrie et des données de puits. Thèse de Doctorat, Université de MontpellierII, pp250 Takherist&Hamdi, 1992, flux de chaleur dans les bassins Nord Sahariens Ziegler P.A., 1989. Evolution of the north atlantic – An overview. In Tankard A.J. and Balkwill H.R. (ed.), Extensional tectonics and stratigraphy of the north atlantic margins. AAPG Mem. 46, p. 111 – 129.

9

Fig1: Geological setting of the Algerian south western basins

Fig 2. : Stratigraphic colon in the western sedimentary basins.

10

Non interpreted section a

Interpreted section b Fig10: Seismic line showing a thrust fault in the carboniferous (frontier basin)

gas flow in the X direction gas flow in the Y direction

b

Fig18: geometrical modelling and gas flow direction

a: 3D network

c: gas flow direction

Fractures network and simulation of the gas flow Direction. The calculation of the spacing between fractures also allowed adjusting the fractures network aatt tthhee llooccaall ssccaallee ((aa ffeeww tteenn ooff mmeetteerrss)) and to determine the direction of the gas flow around the wells. This simulation showed the importance of fractures NW-SE and joints N100 and their contribution to the gas flow.

a

11

Fig. 19: 3D porosities model of' the upper Ordovician unit