06_Szilagy.indd 77 Sebők Szilágyi and Geiger: Sedimentological study of the Szőreg-1 reservoir (Algyő Field, Hungary)... � AB STRA CT The paper presents a 3D sedimentological characterization of a delta plain reservoir. The analysis started with de- tailed core description studies focusing on the genetic interpretations of the sedimentary structures and multivariate statistical analyses (including Q-mode cluster and R-mode factor analysis), of the sporadic grain-size distributions. These results made the depositional processes of channels, bar-crests and crevasse splays probable. By calibration of the well log responses to the cores with total core recovery, we were able to identify two types of vertical facies tran- sitions: (1) mouth-bars and channels; (2) crevasse splays, crevasse channels and minor mouth bars. In addition, a vertical sequence of a relative thick distributary mouth bar was also recognized. Lateral extension was determined by contouring the sand content on surfaces parallel to the upper time horizon. On these surfaces, the sand content and porosity grid systems facilitated the use of an unconventional 3D geo-cellular modelling approach for both well-log- porosities and sand contents. Since vertical depositional facies were extended into the 3D, rock bodies of the most characteristic channel and mouth bar environments could be cut out from the porosity and sand content models. In this way not only their general geometry and internal characteristics, but also the basic geostatistical features (ratio of anisotropy, the direction of major and minor continuity) of their sand content and porosity were identifi ed. Keywords: delta environment, 3D model, distributary channel, mouth bar Sedimentological study of the Szőreg-1 reservoir (Algyő Field, Hungary): a combination of traditional and 3D sedimentological approaches � Szilvia Sebők Szilágyi1 and János Geiger2 1 MOL Plc, Hungary, Exploration & Production; (szszilagyine@mol.hu) 2 University of Szeged, Department of Geology and Paleontology, H-6722, Szeged, Egyetem u. 2., Hungary; (matska@geo.u-szeged.hu) doi: 104154/gc.2012.06 Geologia Croatica 65/1 77–90 19 Figs. Zagreb 2012 Geologia CroaticaGeologia Croatica 1. INTRODUCTION The studied area is located in Central Europe, in the Pan- nonian basin, South-East Hungary. The Algyő Field is the largest Hungarian hydrocarbon accumulation consisting of several oil and gas bearing reservoirs (Fig. 1). A comprehen- sive description of the structural and geological setting of the fi eld is given in MAGYAR et al. (2006). The upper members of these reservoirs have developed in a Pannonian delta system due to complex lateral shifting and prograding phases. They can be subdivided into slope (Algyő Formation) and delta plain (Újfalui Formation) rock bodies. The sills of the individual reservoirs were formed during the delta abandonment phases. Consequently, they can be accepted as time horizons in modelling studies. Be- low these series, the lower reservoirs are regarded as turbi- dity rock bodies (Szolnok Formation) of partly prodelta fans and partly deep basin origin (RÉVÉSZ, 1982). Besides the evident economic importance, the rock bod- ies comprising the Algyő Field can be considered as an al- most complete sequence of general Pannonian (s. l) basin fi lling accumulation (BÉRCZI & PHILLIPS, 1985; BÉRCZI, 1988). Here, according to the Hungarian Neogene geochro- nological terminology we use ‘Pannonian (s. l)’ for Late Mi- ocene-, Pliocene sequences. In the Algyő Field, the Pannon- Geologia Croatica 65/1Geologia Croatica 78 ian delta progradation is one of the youngest within the Great Hungarian Plain. On the basis of seismic data, VAKARCS et al. (1994) then VAKARCS (1997) published the fi rst re- sults of mapping the shelf-margins of the prograding basin fi lling delta-systems. The most recent results have shown that self-margins reached the vicinity of the Algyő Filed about 5.7 million years ago (MAGYAR, 2010, Fig. 31). The Szőreg-1 oil reservoir with a large gas cap is one of the largest sedimentary successions within the delta plain re- cord (Fig. 1). Its average gross thickness is about 35 m, but locally it can reach 50 m. Earlier work by RÉVÉSZ (1982) and GEIGER et al. (1998) proved its delta plain origin with a signifi cant amount of fl uvial channel sedimentation. The main purposes of this study are to (1) provide a high resolution spatial-temporal description of the sedimentolo- gical evolution of the Szőreg-1 reservoir rock body and (2) reveal the real (not simulated) 3D geometry of some depo- sitional subenvironments characteristic of the delta-plain, using 3D contouring of quantitative well-log properties. Basic aspects of the geo-cellular 3D modelling, methods of well-log calibration, multivariable classifi cations and in- terpretations of the grain size data have already been pub- lished (e.g. PARK, 1974; JORESKOG, 1976; GEIGER, 1986; GEIGER & WEROWSKI, 1992, MOLNÁR & GEI- GER, 1995, GEIGER & KOMLÓSI, 1995, GEIGER, 2002; 2004). In this study we emphasize those aspects that can play a signifi cant role in the interpretation of new data. 2. ROCK TYPES OF SZŐREG-1 According to the well logs the internal heterogeneity of Szőreg-1 is considerable (Fig. 2). On the basis of sporadic cores, the following general picture can be compiled concerning the sedimentary rocks of the reservoir. The general succession starts with argilla- ceous marls, which can alternate with claystone, marl, and coaly argillaceous marl. Their average effective porosity is about 0.07 (7%). Argillaceous marls are also observed at higher stratigraphic positions, too, where they are interbed- Figure 1: Theoretical cross section of the Algyő Field. Figure 2: Geological cross section result- ing from the 3D lithological model of the Szőreg-1 reservoir. Geologia Croatica 79 Sebők Szilágyi and Geiger: Sedimentological study of the Szőreg-1 reservoir (Algyő Field, Hungary)... ded with lignite bands. In this latter situation, the vertical sequence of argillaceous marl, marl and lignite transitions are very characteristic. The thickness of the lignite bands or layers can vary from a few centimetres to tens of centimetres. Foliation and shiny sliding surfaces are important features. This lower succession expresses abandonment processes of the previous delta lobes, as derived from the general sed- imentological evolution of the Algyő Field (RÉVÉSZ, 1982, GEIGER 2002, 2004). Fine siltstones are grey – dark grey rocks showing fo- liation or earthy cracking. Their average effective porosity is about 0.1 (10%). They contain a great amount of macro- fauna, in some places even resembling lumashell too. This rock type is generally rich in coalifi ed plant fragments. Fine siltstones show gradual coarsening upward (CU) transition toward coarse siltstones, but they can also be interbedded with thin fi ne sandstones. Coarse siltstones are grey with frequently ochre discol- ouration. Their average effective porosity is about 0.22 (22%). They contain quite high amounts of mica and fi ne sand. In general, coarse siltstones are laminated, picked out with coalifi ed plant fragments. Fine sandstones are light grey, and similarly laminated to the coarse siltstones by coalifi ed plant fragments and mica. They usually represent gradual transitional stages between coarse sandstones and coarse siltstones. The average effec- tive porosity is about 0.28 (28%). A yellowish gray colour characterizes coarse sandstones. They can also be laminated, but a massive, structureless de- velopment is also common. In some places, the clay mineral content of the pore space can be high, due to the presence of syngenetic clay and carbonate minerals. Carbonate also oc- curs as carbonate grains. Horizontal bedding, and high and low angle cross bedding/laminations are the most character- istic sedimentary structures. The average porosity is 0.30– 0.32 (30–32%). Rock types, with the exception of the underlying argil- laceous marl, do not form laterally continuous layers. They can be followed to only for tens or hundreds of metres. In- terfi ngering of the different rock types is a typical character- istic almost everywhere within the studied section (Fig. 2). 3. METHODS A full-scale relational, sedimentological data base has been developed which includes qualitative and quantitative data about core analyses, horizontal layering, facies identifi ca- tion, and well log interpretation. For macro sedimentologi- cal analysis it contains 382 grain size analyses from 53 wells, 291 routine petrophysical analyses from 356 wells, and 156 Hg-capillarity measurements in 126 wells. Geophysical logs and their quantitative petrophysical interpretations (at 0.2 m intervals) are available from 512 wells penetrating the whole reservoir thickness. They were used for mega-scale sedimen- tological analysis including study of well-log responses, and 3D modelling. 3.1. Detailed core descriptions and facies identifi cations The focus was on genetic interpretations of the sedimentary structures and textures, in order to outline both the main depositional processes and the sedimentary facies together with some vertical core profi les. Core descriptions played a signifi cant role in the identifi cation of depositional facies in three key-wells. Descriptions are based on the system sug- gested by MIALL (1966, 1999). In identifi cation of deposi- tional facies of the delta plain, and delta front succession, publications by REINECK & SINGH (1973), GALLOWAY (1975), WRIGHT (1977), POSTAMA (1995) were useful, together with some recently published Pliocene, Pleistocene and Holocene examples (e.g. HOOGENDOORN et al., 2005, LONGHITANO, 2008; ROHAIN et al., 2008, HOO- GEN DOORN et al., 2008) 3.2. Multivariate statistical analyses (including Q- mode cluster and R-mode factor analysis) of the sporadic grain-size distributions According to some earlier works, the lateral variability of the transporting-depositing media can be taken into account by lateral extension of genetic groups determined via cluster analysis of grain-size distributions (PARK, 1974; GEIGER, 1986; GEIGER & WEROWSKI, 1992). For this purpose 382 grain-size distributions have been classifi ed using an r-hierarchical, Q-type clustering tech- nique. In this analysis the similarity coeffi cient was the co- sine of the angle between any two vectors representing the samples. The method for reduction was the centroid ap- proach (MOLNÁR & GEIGER, 1995). Sedimentological interpretation of the resulting discrete groups and subgroups was carried out by the combined use of R-mode factor analysis (JORESKOG, 1976), and interpre- tation of the identifi ed sedimentary structures. The R-mode factor analysis was applied in a cluster-by-cluster manner to the grain size constituents, with the Md and C value of sam- ples belonging to a particular cluster (JORESKOG, 1976, MOLNÁR & GEIGER, 1995). 3.3. Calibration of well log responses to cores with total core recovery In this phase we used the method introduced in Hungary by RÉVÉSZ (1980), where Microlog, SP and Gamma logs were used for calibration. The resulting “electrolithological units” were the tools used to extrapolate the rock types to the un- cored regions. Beside this feature, a semi-quantitative mea- sure, the so-called sand-content (of 1m thick units), was de- rived from the types of electrolithologies (Fig. 3, RÉVÉSZ, 1980). This is the sandstone content (in percent) of the rock body with unit cross-sectional area and 1m thick (Fig. 3). For instance, 80% sand content means that the 80% of the total rock volume can be expected to be sandstone, but it is not necessarily a continuous sandstone layer. This pa- rameter can refl ect both the existence and the intensity of the palaeocurrents (RÉVÉSZ, 1980). Geologia Croatica 65/1Geologia Croatica 80 3.4. 3D modelling approach and high resolution facies analysis The huge number of wells penetrating the reservoir facili- tated the application of a special geo-cellular 3D modelling approach (GEIGER, 2002). Quantitative geophysical inter- pretations were available at 0.2 m intervals in each well. Ap- plying any vertical interpolation in building 3D models would have smoothed out the small scale vertical variances which are needed for analysis by high resolution stratigra- phy. The main goal was to emphasize the stratiform charac- teristics of the depositional systems within the reservoir thickness. The easiest method is to (1) cut the reservoir rock body by surfaces parallel with the bottom of the sill; (2) per- form lateral parameter estimations on these surfaces; (3) build a 3D model by putting the grid systems of the second step below each other (Fig. 4). If there is no considerable lateral erosion between the top and the bottom of the reser- voir, these grid systems should honour the temporal evolu- tion (HOULDING,1994). On each cutting surface, the data locations can be iden- tifi ed by crossing the well trajectories and the cutting sur- faces. Between any two cutting surfaces, the petrophysical data (coming from well-log interpretations), of the 0.5 m thick vertical intervals were given by the average of the da- tum falling to these intervals. These average values were as- signed to data locations on the cutting surfaces. The algo- rithm of lateral extension of porosity and sand content readings was the ordinary kriging with linear semi-vario- gram model. The resulting grid systems can be used in two different ways. On each cutting surface, the grid systems are of the same size. The easting and northing coordinates of each grid node are known. The distance of each nodal point from the topmost surface is also known. This distance is uniformly 0 m on the fi rst cutting surface, 0.5 m on the second one, etc. Note, that in this way we applied the so called stratigraphi- cal coordinate system introduced by DEUTSCH (2002). In addition, certain petrophysical values (from the lateral ex- tensions) are also assigned to each grid node. So, there is everything at hand to visualize a geo-cellular 3D model of the rock body (HOULDING, 1994). This is a trivial applica- tion of the grid systems. The grids defi ned in a vertical sequence, however, can play a signifi cant role in the high resolution facies analysis of the key wells, as well. In these wells the depositional facies of the reservoir section are known. That is, at the crossing of a key well trajectory and a particular cutting surface, the depositional facies is already known from the previous steps. The task is to Figure 3: Electrolithological types and their sand contents defi ned by mi- crologs and cores with total core recovery (after RÉVÉSZ, 1980). Figure 4: 3D geo-cellular modelling process (after GEIGER, 2002). Geologia Croatica 81 Sebők Szilágyi and Geiger: Sedimentological study of the Szőreg-1 reservoir (Algyő Field, Hungary)... fi nd that particular sand contour on this surface which coin- cides with the geometry of the depositional facies identifi ed (Fig. 5). In this way, after drawing the results of facies analysis on the maps, the boundaries of the different depositional envi- ronments can be defi ned by sand isopachs. This method is called high resolution facies analysis, since it gives the 3D ge- ometry of any depositional environment with 0.5 m of vertical resolution (GEIGER & KOMLOSI, 1996, GEIGER, 2002). This approach has two constraints. Firstly, the topmost surface of this sectioning process should be a time horizon. Secondly, there must not be any tectonic movements, the ef- fects of which are laterally variable. The sill of the Szőreg-1 reservoir was formed during a delta abandonment phase. Therefore it can be considered as a time horizon in modelling studies. Moreover, at the time of deposition, the top argillaceous marl can be regarded as a quasi-horizontal surface. The seismic interpretations have proven that within the geological time frame of the Szőreg-1 reservoir there were no lateral tectonic movements. So these stacked maps/grids will refl ect the depositional history in the time interval of the accumulation of the Szőreg-1 reservoir. 4. RESULTS 4.1. Sedimentary structures In both the sandstone and coarse siltstone series, the most frequent bedding form is horizontal planar lamination (Fig. 6, A). In coarse siltstones it represents undisturbed sedimen- tation following Stokes’ law, but in sandstones it can be in- terpreted as a planar bedding in an upper fl ow regime. The latter proves the existence of strong sand laden currents. Characteristically, the planar beds of the upper fl ow regime are associated with trough cross bedding (Fig. 6, B). This feature may express channel progradation, which is a natural phenomenon in the course of delta accumulation both in mouth bar evolution, and crevassing. Fine siltstone and argillaceous marl series often have small sand lenses which can be derived from temporary sandy accumulation from weak currents within interdistrib- utary bays (Fig. 6, C). Fine siltstones and argillaceous marls exhibit character- istic bioturbation. Load structures and convolutions in the Figure 6: Characteristic sedimentary structures from the Szőreg-1 reservoir. Figure 5: High resolution facies analysis (after GEIGER & KOMLOSI, 1996; GEIGER, 2002). Geologia Croatica 65/1Geologia Croatica 82 form of soft sediment deformation are also very frequent (Fig. 6, D). 4.2. Statistical analysis of grain size distributions The sedimentological interpretations of the discrete groups and subgroups were carried out by combined use of R-mode factor analysis and the interpretation of the identifi ed sedimen- tary structures. The R-mode factor analysis has been applied in a cluster-by-cluster manner for the grain size constituents, and the Md and C value of grain size distri butions. As a result the samples could be subdivided into two main groups. They are sandy coarse siltstones and fi ne- co- arse grained sandstones (Fig. 7). The sand:silt ratio of sam- ples belonging to the fi rst group is about 1. In this group, beside the fi ner parts of active channels, interdistributary bays and crevasse splays, the sediments of distributary chan- nels were also identifi able (Fig. 7). The second group (fi ne-coarse sandstones) represents the coarse material deposited as channel-lag of the active distributaries and the coarser grained sediments of distribu- tary mouth bars (Fig. 7). To study the relationship between grain size distribution and sedimentary facies, the core survey data were genetically divided into fi ve subfacies and 1 transitional subfacies. These include (1) natural levee and channel complexes, (2) distrib- utary channels, (3) crevasse splays, (4) prograding crevasse splays, (5) deposits of swamp and channel abandonments, (6) crevasse and abandonment as a transitional subfacies Their averaged cumulative grain size distribution curves are shown in Fig. 8. Figure 7: Results of the multivariate analysis of grain si ze distributions. Figure 8: Averaged grain size distributions by depositional facies. Geologia Croatica 83 Sebők Szilágyi and Geiger: Sedimentological study of the Szőreg-1 reservoir (Algyő Field, Hungary)... The swamp-channel abandonments are the most hetero- geneous in the light of grain size distribution. More than 50 % of the sediments are composed of grains smaller than 0.02 mm (fi ne silt and fi ner sediments). Sediments with grain-size diameters greater than 0.1 mm are subordinate (Fig. 8). Crevasse splays are composed of a mixture of grain sizes from two origins. They represent distributary channels, and the remains of natural levees that were occasionally eroded. These sediments were deposited under high transportation energy of fl ood events. The presence of a coarse grain size and the relatively weak sorting also refl ects these processes in the cumulative grain size distribution. The recurrence of the processes characterized above, can result in the progradation of crevasse splays. The outcome of these events is a slightly better sorting texture than that of distributary channels. More, than 50 % of the grains are be- tween 0.1–0.3 mm. This population may have originated from two or three sources: (1) reworked and redeposited ma- terial of former crevasse splays, (2) the results of channeli- zation or avulsion, and (3) sediments of natural levees. It has the coarsest grains, which is the result of the higher trans- porting energy. In contrast, sediments of prograding crevasse splays are the most well sorted, and have the most sloping grain size curves. The frequency of grains smaller than 0.01 mm is around 10 % (Fig. 8). The average cumulative grain size distribution of the distributary channels is similar to that of prograding crevasse splays, but the coarsest grain size is slightly smaller. They are well sorted, with a modal grain-size of fi ne sand, where 40 % of the grains are between 0.1–0.2 mm (Fig. 8). The natural levee – channel complexes are characterized as a transitional subfacies. The shape of the individual, not averaged, curves can be informative in determining whether it results from an early or late channelizing state of natural levees. The greatest grain size and the modal grain size class are affected by the greatest and the average energy of sedi- ment transport. Based on the cumulative grain size diagrams of the studied subfacies, the following succession can be es- tablished according to decreasing fl ow velocity: (1) prograd- ing crevasse splays, (2) crevasse channels incised to the de- posits of crevasse splays, (3) crevasse splays, (4) natural levee- channel complexes, (5) deposits of late crevassing and abandonment processes, (6) deposits of swamp and channel abandonments (Fig. 8). 4.3. Depositional facies The identifi cation of depositional facies resulted from hydro- dynamic interpretations of sedimentary structures and grain size distributions. Lateral extension of facies away from the vertical columns determined in the wells is demonstrated by the maps of the corresponding mapping surfaces. 4.3.1. Minor mouth bar – channel transition This transition can be detected in different stratigraphic po- sitions within the reservoir. It can develop when a linear fl ow Figure 9: Vertical transition of a mouth bar and channel facies. Figure 10: Vertical transitions of a crevasse splay and channel facies. Geologia Croatica 65/1Geologia Croatica 84 enters a still water bay. In this situation the sediments of the prograding fl ow are initially coarse siltstones (Fig. 9, A) which are followed by gradually coarsening sandstones (Fig. 9, B). On Figs. 9A & B, the still water sedimentation, and ge- ometry of the outer bar can clearly be identifi ed. Later, in front of the channel head, a bar-crest develops as the result of channel and basin interactions. During this process the accumulation progrades toward the basin or bay. On Fig. 9C, this minor mouth bar appears as a characteristic kidney- shaped geometry. The mouth-crest with high sand content can also be identifi ed. In the next phase the distributary chan- nel cuts to the body of the minor mouth bar. 4.3.2. Crevasse splay – Crevasse channel – Minor mouth bar transition One of the typical examples of crevasse splays is shown on Fig. 10A. The upper part of the core may be identifi ed as a vertical section of a mouth bar. The lateral extension of this facies represents a small channel ‘breaking out’ from the mi- nor bar with a NW–SE depositional strike (Fig. 10, C). 4.3.3. Minor mouth bar facies In some situations the bar facies appears as a ‘stand-alone’ succession, as in Fig. 11. In addition to the prograding se- quence, (upward coarsening grain size tendency), the geom- etry also shows how bar crest development follows outer bar sedimentation. 4.4. Sand isopach and facies distribution maps The modelling process of the rock body was extended to 45 m below the permeable top. This volume was sectioned by surfaces being parallel to the top. The vertical distance be- tween any two surfaces was 0.5 m. Their lateral resolution was 100 x 100 metres. The following development can be summarized for the rock body from these maps. From 40 to 45 m below the top the sedimentological evolution is characterized by quite a large distributary mouth bar developed at the south-western rim of the reservoir (Fig. 12, A). The area of this mouth bar is about 14 km2, its thick- ness is about 5 metres. This rock body gradually prograded into a shallow, quiet water bay (Fig. 12, A). Later the focal point of the bay-fi lling accumulation shifted to the North–West in the form of a distributary chan- nel. This channel is fl anked by crevasse splays of different sizes (Fig. 12, B). Its length is about 10–12 km, the average width is 200–300 metres, its thickness is 1–3 metres. This shallow, long channel form is quite characteristic of the re- cent distributary channels. Figure 11: Vertical development of a distributary mouth bar facies. Figure 12: Sand isopach maps on cutting surfaces. Geologia Croatica 85 Sebők Szilágyi and Geiger: Sedimentological study of the Szőreg-1 reservoir (Algyő Field, Hungary)... Within the interval of 19–29 m below the top, the depo- sitional system is characterized by wider lateral extension of a sandy rock body with > 90% sand content. The equivalent sedimentary rocks are coarse and medium grained sand- stones. At that time the focal point of the main depositional processes was in a zone of NE–SW depositional strike (Fig. 12, C & D). The main depositional environments were bi- furcating channels and small distributary mouth bars in front of them (Fig. 12, C and D). The length of the largest bifur- cating channel is about 1.5 km, its width is around 200 me- tres, and the average depth is about 2–3 metres. Sheet like sandstone rock bodies were developed by laterally merging (minor) mouth bars. At this time fi ne siltstone and argilla- ceous marl sedimentation characterizes the NW and SE re- gions of the reservoir. Within the central regions, the largest extent of sand can be detected between 10 and 19 m below the top (Fig. 12, E). From 7 to 10 metres below the top these sandstone bod- ies are disintegrated (Fig. 12, E) and the main depositional system shifted to the NW part. Later in this region, a continuous sandy sheet developed while in the SE foreland the accumulation of siltstones, ar- gillaceous marls and subordinate silty sandstones continued (Fig. 12, G and H). These different identifi ed sandstone bod- ies represent one hydrodynamic system forming one partic- ular reservoir. 4.5. General characterization of the 3D geometry, porosity and sand content of the signifi cant depositional environments As outlined above, one of the most important controls on the whole depositional history is a large distributary channel with SW–NE depositional strike (Fig. 13 A, B). This can be studied between 24–27 m below the top of the reservoir. Based on the 3D representation, it reaches a maximum of 5–6 km in length and about 2–400 m width. The average thickness is 2–2.5 m. The 3D form of this channel partly refl ects its progradation from the SW to the NE, and partly crevassing from the main channel (Fig. 13 A, B). The highest sand content can be seen along the axis of the channel. The lateral and vertical distributions of its po- rosity and sand content are similar to each other. It is also very characteristic, that the largest porosity values belong to the upper part of the channel (Fig. 13, B). Figure 13: 3D geometry and sand content of a large, bifurcating channel. Figure 14: Temporal evolution of a channel system represented in Fig.13 Geologia Croatica 65/1Geologia Croatica 86 The progradation of this channel system is demonstrated in Fig. 14, A–D. In addition to the dendroid channel geometry, it can also be seen that after a progradation phase, during the still-stand period, a mouth bar type geometry with high sand content de- veloped at the hinterland the channel (Fig. 13, B). It suggests the following lateral transitions of the depositional processes: (1) channel progradation, (2) mouth bar evolution, (3) new channel progradation by breaking out from the mouth bar. This sequence is in harmony with the interpretation of both trough cross-bedding and horizontal planar laminations of the core samples coming from a channel environment. The frequency distributions of the vertically averaged nodal sand content and porosity values in the thickness of the channel system show poly-modal histograms (Fig. 15 A and B) with 80.5 %, and 25.5% averages for sand content and porosity respectively. We have performed variography analysis of both average sand content and porosity grids. The results are shown in Fig. 15C and D. The lateral continuity of average sand content is quite high, since the Gaussian- type variogram function gave the best model-result. The main continuity direction of sand content is comparable to the dep- ositional strike of this rock body (Fig. 14, Fig. 15, E). The range in this direction is about 471 m, in the perpendicular direction it is around 339 m. The calculated ratio of anisot- ropy is 0.72 (Fig. 15, E). Since the nugget effect is low, just one-tenth of the sill, this model can characterize almost 90% of the total variance in the terms of linear geostatistics. Figure 15: Statistical and geostatistical characters of sand content and porosity grids of the channel shown in Figs.13 and 14. Signs and abbreviations in parts A and B: SD = Standard deviation. Geologia Croatica 87 Sebők Szilágyi and Geiger: Sedimentological study of the Szőreg-1 reservoir (Algyő Field, Hungary)... In the case of the averaged porosity grid, the main con- tinuity direction is much the same as that of the sand content (Fig. 15, E, F). In practice, this suggests that the textural an- isotropy may either impact on, or determine the spatial var- iance and continuity of porosity. The similarity of anisotropy ratios (0.72 for sand content and 0.81 for porosity) supports this conclusion (Fig. 15, F). The next main physiographic units are the distributary mouth bars. Among them the larger, south-western one, ob- servable between 34 and 39 m below the reservoir top, sig- nifi cantly affected the initial accumulation phase (Fig. 16, A). Based on the 3D geometric representation (Fig. 16, B, C), its planar area is about 14 km2, the thickness is around 5 m. The 3D sand content distribution demonstrates an upward coarsening trend. The highest sand content can be seen on its top showing that the background channel cut itself into the bar crest (Fig. 16, B). The spatial distribution of porosity does not follow the spatial variation of sand content as close as well as in the case of channel system (Fig. 16, C), although the upward increasing effective porosity is typical. Fig. 17 (A to F) demonstrates its temporal evolution. The characteristic kidney-shaped geometry can be clearly recog- nized on each map. On Fig. 17 (C to F), even the bar crest with the highest sand content can also be identifi ed. At the left and right wings of the bar, the development of subaque- ous natural levees refl ects the initial stage of channel cutting and prograding processes (Fig. 17, C). The disturbance of sand contours at the NE border of the bar may represent slid- ing and slumping deformation being frequent within the dis- tal/outer bar region (Fig. 17, D). The frequency distributions of vertically averaged sand content and porosity values exhibit good homogeneity in the case of sand content and poly-modality for the porosity. The average sand content of this bar is 53.7%, the average poros- ity is 19.23%, though the most characteristic porosity values are around 22.4% (Fig. 18, A, B). The lateral continuity of the sand content is also stronger than that of the porosity (Fig. 18, C, D), since a Gaussian model seemed to be the most appropriate for modelling the semivariance of sand content. In the case of porosity, the semivariogram model was an exponential one. For both properties the main continuity directions coincide with the long axis of the bar crest being perpendicular to the direc- tion of bar progradation. The anisotropy ratios are the same Figure 16: 3D geometry and sand content of a mouth bar. Please note: (1) for better visualization the ‘B’ and ‘C’ parts are rotated; (2) although ‘A’, ‘B’ and ‘C’ show the same mouth bar, the ‘B’ and ‘C’ parts are rotated. Figure 17: Temporal evolution of a mouth bar represented in Fig.16. Geologia Croatica 65/1Geologia Croatica 88 (0.78) for both sand content and porosity, (Fig. 18, E, F). Therefore it suggests that the lateral continuity of textural characters had a signifi cant effect on the spatial characters of effective porosity. The 3D sand content model of the Szőreg-1 reservoir indicates that a type of sheet sand drape was developed by lateral merging of the “border” mouth bars. 5. DISCUSSION ON THE DEPOSITIONAL HISTORY OF THE SZŐREG-1 RESERVOIR As a summary, at least three main depositional processes should be taken into account in the evaluation of the depo- sitional history based on traditional sedimentological analy- sis of the core samples, well-log responses, and the 3D mod- elling of sand and porosity content. They are the followings: (1) development of a mouth bar in the south-western part and a large distributary channel breaking out from this body; (2) development of a mouth bar in the opposite direction; (3) progradation of a large distributary channel from the NW to the SE. High resolution analysis has revealed that there were several transport directions from the ‘borders’ to the central parts in the evolution of this rock body. This analysis also highlighted the scale of the most detectable depositional en- vironments. The data and their interpretations have shown that most of the channels were short and several metres in thicknesses. Although the identifi ed mouth bar is relatively Figure 18: Statistical and geostatistical characters of sand content and porosity grids of the mouth bar shown on Figs.16 and 17. Signs and abbreviations in parts A and B: SD = Standard deviation. Geologia Croatica 89 Sebők Szilágyi and Geiger: Sedimentological study of the Szőreg-1 reservoir (Algyő Field, Hungary)... As a consequence, the above mechanism resulted in the infi lling of an interdistributary bay mimics larger delta sedi- mentation. The main difference is obviously the scale and the fact that the driving force, in the accumulation of Szőreg -1, was a multiple source crevassing process and the main delta channels were outside of the study area. According to the depositional strikes they must have been at the SW and NE wing of the Algyő Field. 3D characterization of the revealed channel and minor mouth bar environments also suggests a particular depositional environment where the coincidence of lateral continuity meas- urements (ratio of anisotropy and the main continuity direc- tion) of sand content and porosity, shows that depositional processes have left a signifi cant fi ngerprint on the lateral dis- tributions of porosity. It is important to remember that these properties have been determined independently, since the sand content came from qualitative and porosity from quan- titative well log interpretations. One of the most trivial in- terpretations of this result is that the defi nition of fl ow units in a dynamic reservoir simulation should rely on the corre- lation of those steps of sedimentological history which were outlined in Section 4.5. ACKNOWLEDGEMENT The authors would like to express their acknowledgement to the Hungarian Oil and Gas Company (MOL Group) for per- mission to publish these results. Also, many thanks for the two reviewers for their valuable suggestions and comments. REFERENCES BÉRCZI, I. & PHILLIPS, R.L.(1985): Processes and depositional envi- ronments within Neogene deltaic-lacustrine sediments, Pannonian Basin, Southeast Hungary.– Eötvös Lóránd Geophysical Institute, Geophysical Transactions, 31/1–3, 55–74. BÉRCZI, I. (1988): Preliminary Sedimentological Investigations of a Neogene Depression in the Great Hungarian Plain.– In: ROYDEN, L.H. & HORVÁTH, F. (eds): The Pannonia Basin: a Study in Basin Evolution. AAPG Memoir, 45, 107–116. COLEMAN, J.M., CAGLIANO, S.M. & WEBB, J.E. (1964): Minor sedimentary structures in a prograding distributary.– Mar. Geol., 1, 240–258. ELLIOTT, T. (1978): Deltas.– In: READING, H.G. (ed): Sedimentary En- vironments and Facies. Blackwell Scientifi c Publications, 105–107. GALLOWAY, W.E. (1975): Process framework for describing the mor- phologic and stratigraphic evolution of deltaic depositional sys- tems.– In: BROUSSARD, M.L. (ed.): Deltas, Models of Explora- tion. Houston Geological Society, 87–98. GEIGER, J. (1986): Üledékes homokkőtestek szöveti és morfogenetikai vizsgálata [Textural and morphogenetic study of sedimentary sand- stone bodies – in Hungarian].– Földtani Közlöny, 116/3, 249–266. GEIGER, J. (2002): A pannóniai Újfalui (Törteli) formációban levő Algyő-delta fejlődéstörténete: I. Az Algyő-delta alkörnyezeteinek 3D modellezése [Depositional history of Algyő-delta (Újfalui (Törteli) Formation: 3D modelling of the subfacies of Algyő-delta – in Hungarian].– Földtani Közlöny, 133/1, 91–112. GEIGER, J., KISSNÉ, V.K. & KOMLÓSI, J. (1998): A Szőreg-1 telep 3D rezervoár geológiai modellje [3D reservoir geological modell of Szőreg-1 reservoir – in Hungarian].– Open-fi le report of MOL Engineering Offi ce of Research and Development, #156-5636, 216 p. large in its planar area, the recognized thickness is not larger than 5 metres. These scales are not comparable with those of well-developed delta systems which may act in the Pan- nonian accumulation of the Great Hungarian Plain. The ev- idence, however, shows a real delta succession. This contra- diction can be solved by supposing that this rock body is the result of processes infi lling a larger interdistributary bay. Some well documented recent examples (e.g. COLOMAN et al., 1964, ELLIOTT, 1978) have shown that in these cases crevassing is the main driving mechanism in the accumula- tion of interdistributary bays. In this situation, the stabilized crevasse channels with minor mouth bars developed in front of them, can act as delta lobes. The infi lling process is achieved by their lateral accretion (ELLIOTT, 1978 p. 107). According to our results it can be concluded, that quite a large amount of interdistributary bay sedimentation should be taken into account within the study area (Fig. 19, A). Dur- ing the initial stage, the sandy accumulation concentrated in the SW part, where the appearance of one larger and several smaller distributary mouth bars expressed the preliminary lobe-prograding period (Fig. 19, B). Later these discrete mouth bar bodies merged and a laterally extended sheet-like sand body was formed within the SW region (Fig. 19, C). As a background main channel or delta lobe become more ac- tive, new higher energy distributaries cut down into this sheet (Fig. 19, D). By their lateral migration a new generation of sheet-like sand bodies form and progrades toward the deeper parts of the bay. In front of the channels, new mouth bars develop and merge, meanwhile in the hinterland a typical fl uvial deposition dominates (Fig. 19, E). Figure 19: Depositional history of the Szőreg-1 reservoir. Geologia Croatica 65/1Geologia Croatica 90 GEIGER, J. & KOMLÓSI, J. (1995): 3-D geological simulation of a del- taic reservoir in Hungary.– AAPG Congress, 1995, Nice, Proceed- ings, 65–80. GEIGER, J. & KOMLÓSI, J. (1996): Szedimentológiai, geomatematikai 3-D modellező rendszer törmelékes CH-tárolókban [Sedimentolog- ical, geomathematical modelling system in clastic reservoirs – in Hungarian].– Kőolaj és Földgáz, 2, 53–81. GEIGER, J. & WEROWSKI, P.V. (1992): Geomathematical modelling of a turbiditic reservoir in SE. Hungary.– In: Amilcar Soares (ed.): Geostatistics Troia’92, 1, 311–323, Kluwer Academic Publishers, Dordrecht–Boston–London. HOOGENDOORN, R.M., BOELS, J.F., KROONENBERG, S.B., SIMMONS, M.D., ALIYEVA, E., BABAZADEH, A.D., & HUSEYNOV, D. (2005): Development of the Kura delta, Azer- baijan: a record of Holocene Caspian Sea-level changes.– Mar. Geol., 222/223, 359–380. doi:10.1016/j.margeo.2005.06.007 HOOGENDOORN, R.M., OVEREEM, I. & STORMS, J.E.A. (2008): Process-response modelling of fl uvio-deltaic stratigraphy.– Com- put. Geosci., 34, 1394–1416. doi:10.1016/j.cageo.2008.02.006 JORESKOG, K.G. (1975): Geological factor analysis (Methods in Geo- mathematics, Vol. 1).– Elsevier, 190 p. LONGHITANO, S.G. (2008): Sedimentary facies and sequence stratig- raphy of coarse-grained Gilbert-type deltas within the Pliocene thrust-top Potenza Basin (Southern Apennines, Italy).– Sediment. Geol., 210, 87–110. doi:10.1016/j.sedgeo.2008.07.004 MAGYAR, I. (2010): A Pannon medence ősföldrajza és környezeti viszonyai a késő miocénben [Paleogeographical and paleoenvi- ronmentological characters of the Pannonian basin in the Late Miocene – in Hungarian].– GeoLitera, Szeged, 140 p. MAGYAR, I., FOGARASI, A., VAKARCS, G., BUKÓ, L. & TARI, C., G. (2006): The largest hydrocarbon fi eld discovered to date in Hun- gary: Algyo.– In: GOLONKA, J. & PICHA, F.J. (eds.): The Car- pathians and their foreland: geology and hydrocarbon resources. AAPG Memoir, 84, 619–632 p. MIALL, A.D. (1966): The geology of fl uvial deposits.– Springer, 582 p. MIALL, A.D. (1999): In defense of facies classifi cations and models.– J. Sediment. Res., 69/1, 2–5. MOLNÁR, B. & GEIGER, J. (1995): Possibility for subdividing appar- ently homogeneous depositional sequences by combined use of sed- imentological, paleontological and mathemathical methods.– Geo- Journal., 36, 2/3, 169–177. POSTMA, G. (1995): Causes of architectural variations in deltas.– In: OTI, M.N. & POSTMA, G. (eds.): Geology of Deltas. Balkema, Rotterdam, 3–16. REINECK, H.E. & SINGH, I.B. (1973): Depositional sedimentary en- vironments.– Springer Verlag, New York, 103–142. RÉVÉSZ, I. (1980): Az Algyő-2 telep földtani felépítése, üledékföld- tani heterogenitása és ősföldrajzi viszonyai [Hydrocarbon deposits Algyő -2: geological structure, sedimentological heterogeneity and paleogeographic features – in Hungarian].– Földtani Közlöny, 110, 512–539. ROHAIS, S., ESCHARD, R. & GUILLOCHEAU, F. (2008): Deposi- tional model and stratigraphic architecture of rift climax Gilbert- type fan deltas (Gulf of Corinth, Greece).– Sediment. Geol., 210, 132–145. doi:10.1016/j.sedgeo.2008.08.001 VAKARCS, G., VAIL, P.R., TARI, G., POGÁCSÁS, GY., MATTICK, R.E., & SZABÓ, A. (1994): Third-order Middle Miocene-Early Pliocene depositional sequences in the prograding delta complex of the Pannonian basin.– Tectonophysics, 240, 81–106. VAKARCS, G. (1997): Sequence stratigraphy of the Cenozoic Pannon- ian basin, Hungary.– Unpubl. PhD Thesis, Rice University, Hou- ston, Texas, 514 p. WRIGHT, L.D. (1977): Sediment transport and deposition at river mouths: a synthesis.– Geological Society of America Bulletin, 88, 43–53. Manuscript received June 26, 2011 Revised manuscript accepted December 12, 2011 Available online February 25, 2012