Modelling of maturation, expulsion and accumulation of bacterial methane within Ravneš Member (Pliocene age), Croatia onshore

1 Introduction

The Miocene sediments within the area of research are relatively well explored and with proven hydrocarbon systems of Miocene age. There are a total of 17 oil and gas fields, majority of which are still in the production but with declining reserves [1]. Largest amounts hydrocarbon reserves belong to the oil discovered in the most active exploration period within the Sava Depression (1950 - 1990). The natural gas was considered as a far less worthy resource when compared to oil, but today, such accumulations are explored and considered as a valuable resource [2]. Analogy for the research of Pliocene and Pleistocene sediments, regarding hydrocarbon potential, can be also found in Adriatic gas accumulations. They are of biogene origin, i.e. result of bacterial degradation of organic (dominantly terrestrial origin) matter in a thermally immature stage of hydrocarbon maturation [3]. The type of organic matter in the Croatian part of the Pannonian Basin System Pliocene sediments is similar in composition to the Adriatic ones regarding kerogene (type III and IV) and maturity [4–7]. Lithological composition of Ravneš member and its age can also be correlated to the Northern and Middle Adriatic Pliocene-Pleistocene where poorly consolidated sandstones, sands, marls and clays of thicknesses up to 2000m can be observed [5, 7, 8].

Gas accumulations in the Adriatic hold large reserves [1]. For example, the largest Ivana Gas Field started production with 8 × 10⁹ m³ of recoverable reserves of natural gas [9, 10] and the probability of finding a similar amount of these volumes in the Pannonian Basin System Sava Depression in resembling sediments is a good enough reason for detailed explorations. A preliminary assessment of the generation of bacterial methane from the Ravneš Member was already published [11] but it regarded only the generative potential with no migration and accumulation.

2 Geographical and geological settings

Investigation area is located in the Croatian part of the Pannonian Basin System(CPBS), more accurately Sava Depression. It covers roughly around 340 square kilometres and is located between Ivanic Grad on the West, Novska on the East, Moslavacka gora Mt. on the Northeast and Petrinja on the South (Figure 1).

Figure 1

Outline of the exploration area with basic geology (modified after [6]).

Investigated stratigraphic section belongs to the youngest formation of the Neogene-Quaternary infill of the CPBS - namely Ravneš Member within Lonja Formation. The whole Neogene sequence can be up to 3500mthick in the explored area [12]. It is composed of conglomerate and breccia (pre-Badenian and Lower Badenian), limestone and calcarenite (Badenian, Sarmatian), mostly sandstone with marls (Pannonian and Pontian) and poorly consolidated sandstones, sands and clays (Pliocene and Quaternary) as described in [6, 12–14]. Ravneš Member is part of the Lonja Formation along with five other members (Figure 2). It was selected for this preliminary analysis of hydrocarbon potential because of its lithological composition as it is in some parts almost exclusively composed of coal and in other parts holds large percentage of coal in its total thickness.

Figure 2

Lithostratigraphic subdivision of Lonja Fm. [6].

3 Methods

For this kind of Basin analysis in a preliminary research, there are several steps that need to be made or to be obtained from previous research. Those were: (1) structural modelling, (2) lithology determination (3) organic geochemistry and kinetics definition, (4) calibration. Lithology definition was a result of lithology analysis within 31 well. Part of the process was made using the Schlum-berger Petrel software (structural modelling, lithofacies mapping) while the basin modelling was carried out using the Schlumberger PetroMod software.

3.1 Structural modelling

A subsurface model of Pliocene, Pleistocene and Holocene sediments divided by five e-log markers and borders and into six members was built based on [6]. These e-log markers and borders were defined by standard deviation curves of well log data (spontaneous potential and resistivity curves) and represent a general lithological changes controlled by tectonics and environment influence [15]. The targeted Ravneš Member is located between e-log markers K’ and I’ or border J’ (Figure 2), depending on the location within the depression. Members’ thickness varies from 0 m on the marginal area of the depression to 167 m in its deepest parts. Initial modelling was made in Petrel [6] but for the basin modelling surfaces of each well log markers and the topography were transferred to PetroMod software in which the cell based model was re-built (Figure 3).

Figure 3

Model of the Lonja Formation infill, number s on the axis refer to meters.

3.2 Lithology determination

Initial lithology analysis defined the Ravneš Member as impermeable, shale like dominant facies with substantial amounts of coal (lignite) and organic rich clay [11]. This kind of approach regarded the two types of lithology with transitional mix-type lithology in between. The lithology was controlled [11] by the percentage of coal within the Ravneš Member for wells within the researched area. Also, the Ravneš member was as single vertical cell of variable thickness so the average cell height within the model was from 0 to 167 m.

What was initially disregarded was the fact that there are areas within the member which have substantial amount of sand, especially when the member was layered in three sublayers in equal proportions - Ravneš I - III. Layering was performed with a maximum cell height of 56 m, obtaining more detailed results than in [8]. The lithology of a cell was controlled by either the ratio (>0.25) of permeable and impermeable component (e.g., ss/sh as used in [16]), or by the coal content percentage (%Co) if the lithology was dominantly impermeable (Table 1). If ss/sh was greater than 0.25 than the lithology was defined exclusively by that ss/sh value regardless of the %Co value.

Table 1:

Lithology definition based on ss/sh and %Co values. Mixed lithology numbers represent percentage of each component; Sh-shale or clay; Ss - sandstone or sand; St- silt or siltstone.

Ss/sh	%Co	Lithology
<0.25	0-15	Ravnes_clay
	15-30	Ravnes_Sh80Co20
	30-50	Ravnes_Sh60Co40
	50-70	Ravnes_Sh40Co60
	70-90	Ravnes_Sh20Co80
	>90	Ravnes_coal
0.25-0.40		Ravnes_silt
0.40-0.80		Ravnes_St50Ss50
0.80-1.20		Ravnes_St20Ss80
1.20>		Ravnes_sand

Two mapping techniques were used. First was a relatively classic mapping approach using the Convergent interpolation (CI) algorithm [17]. It obtains relatively smooth transition and is a generally good algorithm for mapping structural surfaces in the subsurface [6]. The maps show relatively smooth transitions (Figure 4a, Figure 5a) depicting a less realistic image when compared to real transition of facies, especially regarding period of the Ravneš Member deposition [6]. Second approach was appliance of the sequential Gaussian simulations (SGS) [18, 19] by which a more realistic map was made with abrupt facies transitions (Figure 4b, Figure 5b)which are characteristic for the Pliocene environments in the Sava Depression [6]. The results were two sets of maps, each containing three maps - one per layer. All are used as facies maps in the basing modelling.Atotal of ten lithology classes were defined (Table 1).

The result of the layering process was a much detailed lithology distribution within the Ravneš Member than previously [11]. It could be clearly observed on cross sections in Figure 6.

Figure 4:

Maps of ss/sh value distribution in Ravneš Member, Ravneš II layer made by CI (a) and SGS (b) algorithms.

Figure 5:

Lithology distribution maps of Ravneš Member, Ravneš II layer made by CI (a) and SGS (b) algorithms.

Figure 6:

Cross section x-x’ (location on Figure 5b) showing the difference in level of detail between a layered Ravneš Member lithology distribution mapped by SGS algorithm (a) and a simple, one layer, lithology distribution by an CI algorithm (b).

3.3 Organic geochemistry properties and kinetics definition

Organic geochemistry data for coals and organic rich clays of the Ravneš Member were obtained from [4, 6]. The coals and clays show dominantly type III gas prone kerogene in a thermally immature stage. At the moment, biogenic reactions should be at the maximum stage regarding expulsion of bacterial methane in the deepest parts of the depression (Figure 7a, b) defined by the transformation ratio (TR) gained from maturity modelling which correspond to the expulsion of bacterial methane in the initial stages of burial [17]. The TR value is presented only for the source rock facies defined in Table 3. Organic geochemistry properties (kerogene type, maturity) of the source rocks are similar to the ones in the Northern Adriatic [5, 7].

Figure 7:

a) Transformation ratio (TR) of the source rock facies by biogenic reactions at present day values on top of Ravneš II layer (SGS), grey areas belong to non-source rock facies; b) cross section showing TR values for Ravneš I-III layers.

Moreover, properties of selected coals from the Ravneš Member are shown in Table 2 along with the several samples obtained from the member in the top and base of the Ravneš Member together with organic rich clays. These properties were then assigned to the Ravneš Member facies defined in PetroMod (Table 3). Algorithm for transformation of the organic matter to hydrocarbons (kinetics) was set for biogenic reactions.

Table 2:

Geochemical properties of coals and organic rich clays from Ravneš and adjacent members [4, 6].

	TOC (%)	HI (mg HC/g Corg)	OI (mg CO2/g TOC)
Sample 1	4.02	64	227
Sample 2	32.15	103	142
Sample 3	39.86	200	137
Sample 4	7.71	71	390
Sample 5	29.1	74	127

Table 3:

Lithology and organic geochemistry definition of the Ravneš Member facies.

Lithology	PetroMod lithology	TOC	Kinetics	HI value	PSE
Ravnes_clay	Shale (typical)	1	Biogenic reaction	70	Source Rock
Ravnes_Sh80Co20	Sh80Co20	5	Biogenic reaction	80	Source Rock
Ravnes_Sh60Co40	Sh60Co40	10	Biogenic reaction	85	Source Rock
Ravnes_Sh40Co60	Sh40Co60	15	Biogenic reaction	90	Source Rock
Ravnes_Sh20Co80	Sh20Co80	20	Biogenic reaction	100	Source Rock
Ravnes_coal	Coal	35	Biogenic reaction	120	Source Rock

3.4 Organic geochemistry properties and kinetics definition

The model output is generated based on all of the data provided in the input. Input data for the model also include “boundary conditions” and paleo geometry. Boundary conditions define the basic energetic conditions for temperature and burial history of the source rock and, consequently, for the maturation of organic matter through time¹. These are surface water interface (SWIT), heat flow (HF) and the paleo water depth (PWD). The data can be altered during each modelling step so the modelled temperature and vitrinite reflectance (VR) correspond to the actual ones in the wells. Heat flow was set at a 70 mW/m² as shown in [20]. SWIT was determined from regional values from [21] and PWD was set according to the reconstructions in [5] in the range from 0 m (recent) to 50 m regarding environments that existed for deposition of members of the Lonja Formation.

Calibration was made only with the fitting of the temperature values (Figure 8), because VR is a poor indicator when dealing with immature source rocks. The model was made for temperature, generation, migration and accumulation for both CI and SGS models.

Figure 8:

Temperature calibration of the model based on well temperature data from well A (location on Figure 7).

4 Results

An estimate of possible generated bacterial methane was obtained for the CI and the SGS maps. Both showed substantial amount of bacterial methane generated - 411.53× 10⁹ m³ for the CI model and 520.56 × 10⁹ m³ for the SGS. The majority of the expulsed volume occurred between 3.5 and 2.5 Ma (Figure 9) after which only small ammounts were expulsed. This was the result of small rate of burial during Quaternary as sedimentation of more than a 100 m thick sediments rarely occurred outside the depocenters, which were at peak expulsion before 2.5 Ma.

Figure 9:

Biomethane expulsion volumes plotted over time showing the period of maximum expulsion for the SGS lithology mapped model.

The TR values are at about 50% for the amount that can be generated in the biogenic reactions domain and the remaining potential is related to the source rocks in the shallow areas (Figure 7). In contrast to large generated volumes of biomethane, the accumulated volumes are rather poor - 0.20 × 10⁹ m³ for the SGS and 0.06 × 10⁹ m³ for CI model. Small accumulation volumes wereexpected within the Ravneš member as it was regarded as dominantly a source rock facies with sporadic occurrences of reservoir facies (Ravnes_St20Ss80 and Ravnes_sand lithology defined in Table 1). There are a large number of accumulations within the SGS model with very small amounts and only a few with volumes of over 1 × 10⁶ m³ (Figure 10). Most of the accumulations, especially of the smaller volumes are a result of stratigraphic trapping due to the lateral lithology change (Figure 10, enlarged area). The volume of these accumulations can be underestimated as the cell grid of the model was relatively coarse (100 × 100 m) in regard to the area of the accumulation (largest one 0.50 km²) according to [22] There are only a few accumulations in the CI model which are a result of a gradual lithology change in the up dip direction and a small number of possible structural traps.

Figure 10:

Small volume accumulations within the Ravneš (III) member made by SGS algorithm. Red pointer shows one of the few larger accumulations (2.58 × 10⁶ m³). Enlarged area shows small volume accumulations due to the stratigraphic trapping.

5 Conclusions

A substantial volume of generated and expulsed bacterial methane is determined by the basin modelling process from the Ravneš Member. Both the CI and SGS lithology approaches gave similar results when volumes of generated methane was regarded - 411.53 × 10⁹ m³ for the CI model and 520.56 × 10⁹ m³ for the SGS. On contrary, volumes of the possible trapped bacterial methane in Ravneš Member are very small and significantly differ - 0.20 × 10⁹ m³ for the SGS and 0.06 × 10⁹ m³ for CI model. The differences in generated and trapped volumes resulted from different lithology distribution between the CI and the SGS derived lithofacies maps.

Although the convergent interpolation provides a safer result (probabilistic approach), the SGS (stochastic approach) is more corresponding to the dynamics and size of the depositional environment during the Lower Pliocene in the Sava Depression. Presented analysis supports the conclusions [6] that Ravneš Member is a good source rock interval for bacterial methane generation, but commercial size accumulations within the member itself are not to be expected except if regarded as secondary targets when drilling for deeper accumulations. Technical problems with inflow of sand into well which can arise from producing in poorly consolidated sands and sandstones, similar to the ones in North Adriatic, should also be taken into consideration when evaluating the possibility of economical accumulations within the Ravneš Member as well as all other accumulations in Lonja Formation.