3.2.2. CO2 Storage Potential in Anticline Structures of the Carbonate Complex

The structural map of the Top carbonate complex horizon in the Croatian part of the Adriatic off-shore (Figure 3) was constructed based on data published in one PhD thesis [28], a Master thesis [24], two graduate theses [55,56], and that in publications in scientific and professional periodicals and proceedings [29,57–59]. The regional structural model is the most important here, so characteristics of structural styles and the location of km-scale structures are described first.

In terms of the predominant structural styles and structures formed in rocks of the carbonate complex and its overlying syntectonic sediments, their orientation and distribution, the three different regions in the Adriatic offshore of Croatia can be distinguished (Figure 3):

1. Area to the WSW of the Premuda, Susak and Lošinj islands, together with the offshore west of Istrian peninsula. In this part of the Adriatic offshore, the top of the carbonate complex horizon gently dips in the WSW direction in a form of monocline (named the North-Adriatic monocline by [30]); in offshore Istria, it represents the gently WSW-dipping limb of the so called Istrian anticline [60,61] or the Istrian swell, sensu [30]. This wide and gentle anticline is bounded to the east-northeast by the frontal thrust of the External Dinarides exposed along the SW margin of the Ci´ ´ carija mountain (Figure 3), while its submerged WSW limb practically continues all the way underneath the submerged thrust front of the Northern Apennines (see in cross-section number 5 in [30]). Thus, the Istrian anticline represents a gently deformed foreland at first for the Dinarides fold-thrust belt during Middle to Late Eocene, and then for the Northern Apennines fold-thrust belt during Late Miocene to Quaternary. In the North-Adriatic monocline, at about 50 km offshore Rovinj where the core of the anticline crops out, a paleogeographic

boundary between the AdCP and the Adriatic basin is nicely preserved. According to [30,62], this boundary is interpreted as the W-dipping Early Jurassic to Paleogene normal fault that is covered by undeformed Plio-Quaternary marls and sands, that, in addition to an absence of instrumentally recorded seismicity along this boundary, suggest that it is at present, tectonically inactive. The same is true of a set of conjugate normal faults found some 20 km east of this boundary and close to the Ivana gas-field (Figure 3).


**Figure 6.** Transversal cross-section 2–2 through the deep well Kate-1, south of the Dugi otok island (after [30]; stratigraphy modified after [15]). Vertical exaggeration is 2:1. Fault planes are marked with black lines. Location in Figures 3 and 4.

Altogether, five structural traps—potential underground CO2 storage objects, are depicted, all of them identified based on the structural map of the top of the carbonate complex. Structures are shown in detail on small maps given in Figure 3. The main characteristics of the potential storage objects are given in Table 3 and main parameters used to calculate storage capacities are given in Table 4. The average effective porosity value has been extrapolated from the laboratory measurements on core samples of the Upper Cretaceous carbonates from a single well in the northern part of the Adriatic offshore [28]. To calculate pore compressibility, the correlation of pore compressibility with net confining pressure developed for carbonate rocks has been used [72]. The correlation itself was developed in Knutson and Bohor [73]. Pore water compressibility was calculated in the same manner as for DSA Dugi otok, using a correlation by Osif [53]. The maximum pore pressure is set based on criteria 90% of 0.2 bar/m fracture gradient; i.e., as 0.18 bar/m [74,75]. In that respect, maximum increase of pore pressure was averaged to 50% from initial pore pressure, which is in accordance with value of maximum pore pressure suggested to be between 1.3 and 1.8 times the initial pore pressure [76]. Storage efficiency coefficient was taken to be 0.05, which is the product of displacement efficiencies (volumetric, *EV* and microscopic, *ED*) and net-to-gross-ratio. For product of volumetric and microscopic displacement efficiencies, the estimated value of P10 in limestone formation storage objects after Goodman et al. [49] amounting to 0.1 was taken, while for the net-to-gross ratio, representing the part of the structurally defined saline aquifer having favourable petrophysical properties needed for CO2 injection (generally corresponding to *Ehn*/*hg* after [49], the value of 0.5 was taken. The value of net-to-gross was practically based on a rule of thumb approach, since there were not enough data to make reliable geological models of these structurally defined aquifers. The porosity data used to calculate pore volume was effective porosity, but it was extrapolated from the neighbouring well, not from the wells drilled-through the structurally defined aquifers. Also, no quantitative data of permeability were available that could be used to assess net-to-gross ratio. Temperatures were estimated using geothermal gradient that was calculated from data on temperature of the sea bottom [54] and the regional isothermal map of formation temperatures at the depth of 3000 m [31]. The calculated values were in agreement with the

geothermal gradient mapped by Jeli´c et al. [77]. Since no data on pressure were publicly available, initial pore pressure was calculated using hydrostatic pressure gradient and this can be regarded as a reasonable assumption; i.e., no overpressure is to be expected, due to the fact that drilling operations encountered problems with total mud loss when entering the carbonate complex [19]. Densities of CO2 were calculated based on the estimated values of pressure and temperature, using equation of state as defined in [46].


**Table 3.** Main characteristics of structurally defined aquifers in carbonates.

**Table 4.** Storage capacity estimation using compressibility method for structurally defined aquifers in carbonates.


\* CO2 density at maximum pore pressure (sum of average initial pore pressure and overpressure caused by injection).

In this way, calculated total storage capacities in five chosen structurally defined aquifers were considerably high, which makes them valid candidates for future exploration activities. Special attention should be given to the fact that in three of five potential storage objects (structures 1, 2 and 5) CO2 is not expected to be in supercritical state, but liquid upon injection, due to low initial average temperatures that are the result of a low geothermal gradient (between 1.2 and 1.5 ◦C/100 m), characteristic for the studied area. This is not necessarily an issue, since according to [78], injecting CO2 in a liquid state is energetically more efficient than in supercritical state, due to its increased density, which results in lower overpressure not only at the wellhead, but also in the reservoir, because a smaller volume of fluid is displaced.

It should be emphasized that the obtained capacities are heavily burdened by the lack of data and subsequent weaknesses of the model used for their calculation and can also be treated as theoretical values only. However, it must be noted that numbers given in Table 4 are more realistic than the estimates given for the Miocene regional aquifer (Table 2), making at least some of these objects targets for future detailed exploration.

#### **4. Discussion**

Trying to estimate the storage capacity in deep saline aquifers (DSA) always disclosed a major problem, because the available data on the subsurface geology are not detailed enough. Even in the mature petroleum provinces deep aquifers were simply not drilled through in many places and there are just a few analyses of their reservoir properties. There are frequent cases where the geometry of the reservoir rock formations can be delineated based on the regional subsurface data, but other parameters—effective thickness, porosity and temperature—need to be extrapolated from the existing hydrocarbon fields in the region, if there are any. This inevitably burdens the storage capacity estimates with a lot of uncertainties. Even more so, knowing that adequate trapping conditions in parts of these regional aquifers will only later be confirmed by targeted surveys. That is why these storage estimates

are regarded as theoretical capacity only (bottom of the techno-economic resource pyramid for the capacity of CO2 geological storage as defined in [79]).

There are the two significantly different types of formations where potential underground CO2 storage objects might be planned and constructed in the Adriatic offshore. Firstly, there are the thick carbonate rock formations ranging in geological age from Triassic to Eocene. On the map of top of carbonate complex (Figure 3) in the more prospective zones (i.e., far from the active faults) altogether, five structures were depicted. Three (2, 3 and 4) of them were drilled by regional wells and no hydrocarbons were discovered, meaning that they can be assessed as structurally defined aquifers. Their main characteristic is the primary and secondary porosity, thus potentially high permeability, which is indicated by total mud loss during the drilling of the mentioned wells [19]. Carbonate rock formations are, in the Adriatic offshore, covered by thick succession of clastic sediments (from Eocene to Holocene age), in which most of the rocks are impermeable, most importantly the Upper Miocene and Lower Pliocene layers. The thickness of the entire clastic basin fill is given in Figure 4. Another interesting potential storage object is the deep saline aquifer—Dugi otok (DSA Dugi otok). This is a regionally defined unit of thick Miocene succession of marls and sands that filled the Dugi otok depression. Looking at the cross-section 2–2 in Figure 6, and given description of regional geology, this regional aquifer might be considered as an object worth the detailed exploration for two reasons—ample impermeable intervals (regional seals) of the Miocene and Pliocene age, and a regional dip SW by one largely undeformed structure, allowing plans to be made for the injection wells on the subsided SW part of the monocline and monitoring wells on the NE side. That is, should such general structure be confirmed by targeted exploration. The drawback for now is in the smaller proportion of permeable layers (estimated net pay of 0.2 is in the Table 2) and the same goes for the true reservoir properties, because they are also only regionally estimated. CO2 storage capacity declared for the DSA Dugi otok is really a preliminary estimate for two reasons—its reservoir rock properties are based only on the data from three wells, and its outline follows the contour 3000 m on the map of thickness of clastic sediments (Figure 4), because that is the area where the Miocene sediments have greatest thickness, and within this area thickness of Pleistocene and Holocene sediments is the greatest, meaning that the Miocene strata are situated in depths exceeding 1000 m. It also has to be noted that the storage efficiency factor is taken to be 0.02 [49], meaning that only such a small proportion of the estimated available pore volume might retain, once being filled with carbon dioxide (at several locations that are still to be found). This storage capacity in aquifers might be prepared for use only after the deliberated exploration of these objects, not only to fully investigate their reservoir properties, but also to confirm the integrity of their cap rocks. The third option, storage objects in the three gas fields might easily be prepared for pilot injections and have upscaling potential, but this will become available only once their reservoirs are depleted and hydrocarbon exploitation licences are expired or terminated.
