Part 3 #GigatonneCCS

Characterizing Aquifers in Data Lean Environments

wave imagewave image

In Part 3 of this series we dive deeper into the Elephant and demonstrate how we extract insight in a data lean environment. 

Following on from the description of the Elephant structure in Part 1 and 2 (using data from PGS19MO2NWS or "The Elephant" dataset) and the encouraging results from the initial analysis of the storage potential, we will look more closely at the stratigraphy within the area of interest, which contains a sequence of up to four early to middle Jurassic sandstone units developed within the subsurface across the wider Norwegian Sea region.

Simplified stratigraphy showing the distribution of aquifers and sealing lithologies. For the Elephant, the seal is the Melke Formation shales and the primary aquifers are within the Garn and Ile Formation sands.

As discussed in Part 2, the Spekk Formation blankets the area creating a reliable regional seal. What is of real interest is the potential aquifers that CO2 could be injected into, with four potential aquifers identified that have been deposited in shallow marine to terrestrial environments:

  • Garn Formation | beach barrier and wave-dominated deltaic deposits
  • Ile Formation | progradational shallow marine shoreface environments 
  • Tilje Formation | highly complex tidal and wave-influenced marine deposits
  • Åre Formation | braided fluvial environment with significant coal swamp development

Each of these units are well known to the hydrocarbon industry as effective reservoirs on the Halten and Dønna Terraces to the west. However, on the Trøndelag Platform, well information is sparse, and so offset well data needs to be considered, alongside information from the seismic analysis. 

Reviewing the Tectonics of the Area to Assess Complexity and Capacity

It is worth considering the tectonostratigraphic development of the area in more detail to appreciate the potential complexity in applying analogues directly from those wells, which, as ever, needs to be done with caution. Most of the wells are within the adjacent prolific oil and gas province developed along the Halten and Dønna Terraces. This region is dominated by rotated fault blocks, the result of intensive polyphase extension impacting the area throughout the late Triassic-Jurassic.  

The terraces are separated from the Trøndelag Platform by the Revfallet-Bremstein-Vingleia fault complex, which fundamentally separated the two areas from each other during the latest Jurassic to lower Cretaceous. However, before this, although the platform and terrace areas possess differing structural styles, deposition across the areas was linked and inferences about the depositional systems and sedimentology on the platform can be made from understanding the regional context of the offset well data. Bunkholt et al. (2021) present an excellent tectonostratigraphic synthesis of this region, drawing on a range of PGS datasets, including PGS19MO2.

Given this challenge, we extracted as much information as possible from the seismic data to tie back with the extensive well information across the Halten Terrace. Below we share a summary of what we have found from the Garn and Ile Formations.

Garn Formation: Waves and Deltas

The Garn Formation is a key petroleum target across the Halten Terrace where extensional faulting has created syn-rift accommodation space and provided ample trapping geometries for conventional exploration. As we have already discussed, the Trøndelag Platform was tectonically quiescent during the main phase of late Triassic and Jurassic deposition, and this helped to give rise to extensive, continuous sandstone-bearing units deposited in fluvial and shallow marine environments. Their continuity and extent is one of the main attractive characteristics of these reservoirs or aquifer systems for a migration-assisted storage concept at scale. We wanted to understand how much we could apply to define our concept using well-based models of the Garn Formation in the literature, which are based on data from the Halten terrace to the west, and apply them to the Elephant. In assessing this offset data, it was important to appreciate the differing structural development between the two areas.

By using high-quality GeoStreamer data we could distinguish the Melke and Not mudrocks from the sandier Garn Formation. Traditional horizon interpretation, where the interpreter attempts to map the top and base proved to be difficult for the Garn Formation due to apparent lateral changes in the sand content within the unit. Therefore, it was decided to apply a more appropriate method for mapping the more sand-prone elements of the Garn, using a probabilistic facies inversion to achieve the best result in isolating the aquifer units and refining our understanding of their distribution. In the figure below you can see the location of areas where the facies inversion predicts a high probability of sandy Garn Formation occurring, displayed in yellow/white with the shales highlighted by red and blue in the overyling unit.  Lateral facies changes are readily apparent, and this compares to a rather different character within the underlying Ile Formation, where the aquifer sandstones are expected to be relatively more continuous than the overlying Garn Formation.


Seismic facies inversion showing the probability of sandstones (white/red is high probability, blue/green is low probability) highlighting the distribution of the Garn and Ile Formations. 

Based on this extraction we wanted to further QC our conceptual understanding of the Garn depositional system by mapping out the distribution across the area of interest. In the figure below, the distribution of the inferred Garn Formation is shown as an isopach map overlain on the Top Melke surface for context with minimum relative acoustic impedance from the interval mapped below it. This extracted information with the sparse local well calibration and models from the data-rich Halten Terrace allowed us to develop a picture of the depositional elements and their distribution across the Elephant to inform the construction of the later reservoir models required for the reservoir simulation.

This Garn Formation depositional element map shows how high-quality GeoStreamer data guide the development of facies maps across the area of interest and highlight key geological features of the aquifers that be directly used in later steps, including reservoir modeling.

Key learnings from this analysis of the Garn formation:

  • With limited local well data to drive our interpretation of the Garn Formation, we use high-quality 3D GeoStreamer data and seismic imaging. We have been able to rapidly extract insight on the reservoir distribution and setting, and offset the relative sparsity of other data, to dramatically improve our understanding of the potential aquifer distribution.
  • The general trend of the depositional system within the area is north-south, which is somewhat at odds with previous basin-scale interpretations. where the provenance and general direction of sediment transport was -east-west, from the direction of what was then the Baltic Shield (modern-day Scandinavia).  We postulate that this might reflect the development of a more complex coastal morpohology locally across the platform  The data clearly improves our understanding of the sedimentary system enabling improved inputs into the corresponding reservoir models. The section below on the Ile Formation clearly illustrates this.
  • Completing interpretation and well ties on the latest generation of 3D data changed our understanding from the earlier well-based concepts.  Tying the wells to the wider survey and regional interpretation in conjunction with the careful scrutiny of the well logs and other well-based data suggests that the Garn has been mis-picked in wells 6507/8-5 and 6507/12-2. Whereas this work indicates the sediments below to the Ile Formation.  This important and rather different system is discussed further below.


Ile Formation: Super-Sized Strand Plains

Prograding strand plain deposits showing across the Ile Formation in section flattened at the base and in plan view defined by spectral decomposition above.

Regionally, the Ile Formation comprises a high-quality reservoir package, and extensive shallow marine sands typify the unit and this is no different in the Elephant based upon the seismic evidence and regional well ties. The package forms a clear low impedance unit that extends across the whole of the Elephant structure, making it a prize target in the hunt for large-scale CO2 storage opportunities and significant pore volume.

The section above shows the Ile Formation flattened at the base and showing clear prograding geometries to the north, again somewhat at odds with the general basin-scale picture for these formations. What we found remarkable is the unmistakable strand plain features which can be observed across the dataset and is displayed in the image above.

In comparison to the Garn Formation in the area, the Ile Formation has been drilled with two cores cut providing excellent ground truth calibration for the seismic observations. The cores contain classic coarsening upwards sequences indicative of lower to upper shoreface facies associations. Wave-dominated structures and processes dominate, with the odd controversial tempestite or storm deposit interpreted for good measure! All in all, classic features of an energy wave-dominated depositional system.

Next Steps: Modeling at All Stages

Given the fantastic geological insights and detail gleaned from the GeoStreamer seismic data combined with the overall structure, we can start to think about how this system might work as a migration-assisted storage concept.  Below our conceptual model of a migration-assisted aquifer has been modified to include key elements from what we have learnt from our interpretation of the Garn and Ile Formation and how these might work together.  

Interestingly we also note that the Garn is ‘plumbed in’ to the top of the Ile in some places, with pronounced evidence of erosion which results in a connection between these two aquifers. How will this work as a migration-assisted system and what might be some of the optimal strategies to develop it as a CO2 store? These are some of the questions that we have started to address with our initial modeling exercise, which has been designed to assess how the various geological characteristics will combine to create an effective large-scale storage complex.

Conceptual sketch illustrating the potential injection strategy, where CO2 is injected into the underlying Ile Formation and allowed to migrate up-dip, where some of it will potentially migrate into the overlying Garn Formation, the shallowest accessible sandstones within the Jurassic aquifer sequence.

To validate these concepts we have been devising and testing different modeling and simulation methodologies throughout the process of developing the Elephant concept. In the next installment, we will share some of the initial simulation results and outline some dynamic considerations relevant to the site. These we have developed in close cooperation with one of our key collaborators, OpenGoSim, to develop our dynamic understanding of plume dynamics and how CO2 could be stored in the site and encourage us to go further with a more sophisticated set of models in the next stage of work.

Get in touch with the authors of this blog, Nick Lee or Bill Powell to find out how PGS can assist you with your CCS characterization challenges.