Update: Reducing CO2 reservoir uncertainty with palynology
Capturing carbon dioxide (CO2) from industrial and power sources and storing it in sub-surface geological formations is an option for reducing emissions into the atmosphere. Palynology (the study of fossil spores and pollen that are common in argillaceous rocks) may have use in the crucial area of the uncertainty caused by geological heterogeneity, specifically the size and arrangement of mudstone baffles to fluid flow. Its use shows how science developed to understand oil and gas extraction can be re-invented to help in the injection and underground management of CO2. Recent work in 2022 and 2023 on Jordanian Permian fluvial clastic sequences has shown how different types of mudstone baffles can be identified by their palynological assemblages. Statistical analysis of palynological assemblages may help in providing an objective, quantitative measure of baffle type and size in the future.
Carbon capture and storage 101
At its most basic, CO2 storage requires that supercritical CO2 be injected into deep subsurface reservoir rocks over a long enough time period to make a difference to atmospheric CO2 concentration. Understanding the behaviour of CO2 in advance of development of these reservoirs is therefore vital to provide confidence for investors and project developers – but also for the governments that will regulate the process and eventually take over the long term liability for a store (Stephenson 2013).
Geological models used so far have aimed at largescale estimates of storage capacity to get a view of the overall feasibility of CCS as a technology, in comparison with other technologies that are on offer. In general these large scale estimates, for example the Energy Technologies Institute UK Storage Appraisal Project (UKSAP) in 2011, tell an encouraging story suggesting abundant storage space that could support a largescale future European CCS industry. However these models are necessarily broad brush and incorporate very limited geological detail, particularly at the sub-seismic (metres to tens of metres) scale. One of the key uncertainties is heterogeneity along the migration path of an injected CO2 plume. These heterogeneities can include those bestowed on the rock by primary sedimentological and paleoenvironmental factors, those from diagenesis and those from tectonic and structural change. Detailed geological data on heterogeneities are often not available for the saline aquifers that are an important target for CCS because their depth falls between the relatively data-rich environments of shallow freshwater aquifers and hydrocarbon reservoirs. Biostratigraphy and palynology can of course help to identify and correlate suitable reservoirs for CO2 storage, but recent research shows that palynology coupled with sedimentological analogue outcrop study can also offer insights and predictive tools to understand heterogeneity.
Analogue outcrop study combined with palynology is not a new thing for understanding reservoirs; oil and gas companies have consistently used outcrop models to better understand their subsurface assets. There are some crucial differences though: the fact of injection rather than extraction, that CO2 is a reactive gas injected in supercritical form, and that CO2 is being placed as far as possible in geological conditions that encourage slow migration, solubility trapping (where CO2 dissolves) and/or stratigraphic trapping.
The famous Sleipner CO2 storage facility is a case in point illustrating the balance that may be sought between injectivity and presence of enough heterogeneity to facilitate solubility trapping. Sleipner is the longest running facility for CO2 storage, having injected around 1 million tonnes of CO2 every year since 1996. A series of repeat seismic surveys has detected layers of rock with high CO2 saturation which show the attenuation of CO2 where it has accumulated below thin sub-seismic mudstone layers. The 4D seismic has also been able to track the rate at which buoyant supercritical CO2 rises through the reservoir, how and when it reached the overlying seal, and how it moved under the seal as more CO2 arrived. Perhaps the most interesting aspect for a stratigrapher is the role of the mudstone layers. Under other circumstances these layers might have been seen as problems for injection – perhaps making less of the reservoir available to injection. But the presence of these baffles has also proved useful in slowing down upward migration allowing more time for reactions to take place dissolving CO2 and leading to more long term storage through carbonation. Thus the low permeability mudstone layers promote ‘solubility trapping’ leaving the overlying seal (the physical trap) to do less of the work of confining the CO2.
Palynology in understanding the size and type of mudstone baffle
The importance of palynology in understanding the geometry of mudstones is a newer discovery. Early indications of this usage came about through the study of Permian mixed mudstone/sandstone fluvial successions during three field campaigns in Jordan. At that time, palynology was being done to support the description of some of the world-class fossil plant discoveries in the Upper Permian Umm Irna Formation, a succession similar to the more fluvial parts of the Sherwood Sandstone Group - a major CCS target in the North Sea.
Fig. 1 the Upper Permian Umm Irna Formation occupies most of the cliff shown (about 500m high). The formation is a useful outcrop analogue for CCS
The Umm Irna Formation is 70m thick and outcrops for approx. 40km north to south along the Jordanian Dead Sea coast (Fig. 1). Detailed field sedimentology allowed the reconstruction of a paleoenvironmental model which indicated broadly the kinds of palaeoenvironment that might lead to argillaceous sediments being laid down amongst the high permeability channel sandstones that dominate much of the formation. These argillaceous sediments could be laid down as baffles to fluid flow in a number of different geometries and lateral and vertical patterns, all of which contribute to permeability heterogeneity in the formation.
During a December 2022 field session, several mudstone units were sampled in great detail to understand not only the general palynological character of different kinds of mudstone baffle, but also how palynology might change within the baffles along their length and through their thickness. The mudstone units sampled fell into three categories:
1. Laterally accreting channels with argillaceous point bar deposits
2. Large abandoned channel fills
3. Small abandoned channels or oxbow fills
These are shown in the context of the palaeoenvironmental model of the Umm Irna Formation postulated by Stephenson and Powell (2013) in Fig. 2.
Fig. 2. Mudstone units sampled in this study: (1) Laterally accreting channels with argillaceous point bar deposits, (2) Large abandoned channel fills, and (3) Small abandoned channels or oxbow fills – all in the context of the palaeoenvieonmental model of the Umm Irna Formation postulated by Stephenson and Powell (2013)
Preliminary results
The first thing to point out is that these are preliminary results from a small number of relatively well defined mudstone units; a number of units are yet to be analysed. The units are at approximately the same level stratigraphically, and previous work on the formation (Stephenson and Powell 2013, 2014) suggests that palynological variability within the formation relates mainly to the position on the palaeo flood plain, rather than evolutionary differences in this relatively short Changhsingian (uppermost Permian, Stephenson and Powell 2014) time interval.
Laterally accreting channels with argillaceous point bar deposits
A spectacular outcrop at the Dyke Plateau locality (see Stephenson and Powell, 2013; 31 32 6 N 35 33 25 E) displays a very large channel (around 50m wide) and the related point bar sediments (Fig. 3). The palaeocurrents and other characteristics of the channel and point bar are described in detail by Stephenson and Powell (2013). The point bar deposits contain a laterally continuous package of argillaceous sediments of various facies that are well known for plant fossils (e.g. Kerp et al. 2021); but also previous sampling (Stephenson and Powell 2013) indicated that the units yielded palynomorphs. Overall the mudstone package is around 70 cm thick on average, and has a lateral extent along the base of the cliff at Dyke Plateau of around 50 m. Fig. 3 (top) shows a tentative reconstruction of the original extent of the argillaceous point bar deposits, taking into account the geometry and size of the channel. Based on this reconstruction, a rough estimate as to the subsurface area of the unit disregarding modern erosion and exhumation would be at least 2500 square metres. This suggests a large baffle to fluid flow between and within relatively high permeability sediments.
In order to characterise the unit both laterally and vertically, it was sampled extensively in four sections, each section containing around 5-7 samples.
Fig. 3. Laterally accreting channel with argillaceous point bar deposits at Dyke Plateau. Upper diagram shows the plan view with a channel to the north, and a tentative reconstruction of the original extent of argillaceous point bar deposits. Lower diagram shows the view from the west with channel to the left and argillaceous point bar deposits in profile. Sections A to D were sampled for palynology (positions approximate).
The palynology of the four sections is remarkably consistent. All the assemblages contain the same dominant species, and these and their proportions tend to remain the same both up through the sections and across the sections. Preliminary statistical analysis bears out the internal consistency of the palynological assemblages from these point bar deposits. Fig. 4 illustrates the palynological composition and stratigraphic trends in one the sections: Dyke Plateau section A.
Fig. 4 Composition and stratigraphic trends in Dyke Plateau section A
Large abandoned channel fills
A large abandoned channel fill was sampled in eight sections from a trench related to a new roadcut along the main road from Amman to Aqaba (31 32 38 N 35 33 27 E). The trench is shown in Fig. 5. The mudstone unit is around 70 cm thick on average and is exposed on the east and west sides of the trench. It also seems to appear in the main roadcut around 30m to the west. This would again suggest a baffle of considerable subsurface areal extent, perhaps around 1000 square metres.
Fig. 5. Large abandoned channel fill exposed in a trench related to a new roadcut
The palynological assemblages in the samples are more internally heterogeneous than those of the Dyke Plateau point bar deposits, with more variation between samples and sections laterally and vertically, and more low-yielding or barren samples. Though overall the palynological samples have similar dominant species and similar proportions to the dominant species of those of Dyke Plateau, preliminary statistical analyses indicate that these assemblages as a whole are significantly different to those of Dyke Plateau.
Small abandoned channels or oxbow fills
Several small argillaceous units were sampled, including a locality south of Dyke Plateau (31 31 57 N 35 31 26 E), consisting of thin and impersistent siltstones within stacked thin sandstone beds, and a locality at Wadi Autun (N31°32‘40.1’’, E35°33’31.7’). Both localities display small argillaceous units that are probably minor channel bases (Fig. 6).
Fig. 6. Left: locality south of Dyke Plateau of thin and impersistent siltstones within stacked thin sandstone beds; right Wadi Autun
Fewer samples could be taken from these units because the argillaceous beds were thinner and less laterally persistent suggesting potential subsurface extents of less than 20 square metres, and more likely around 10 square metres, so only minor baffles to fluid flow.
The palynological signatures of these units are very different to those of the large argillaceous units. They contain different dominant species, lower diversity, and considerable variability from one small argillaceous unit to another. For example the locality south of Dyke Plateau contains common monolete spores and monosaccate pollen. The Wadi Autun locality contains common monolete spores bit also common trisulcate pollen. These spore/pollen groups – particularly monolete spores and trisulcate pollen - are rare in the larger argillaceous units. Statistical tests bear out the high level of difference both between small argillaceous units, and between small argillaceous units and larger units such as major channel fills and point bar deposits.
Preliminary conclusions
So the work up to now indicates the following:
1. Larger baffles (2500 square meters and larger) have diverse palynological assemblages with particular dominant species that tend to be relatively consistent across and through units. Large units from different palaeo floodplain habitats can be distinguished statistically
2. Smaller baffles have lower palynological diversity and different dominant species to the larger baffles, but also different dominant species from one baffle to another
Although this work is incomplete, these early findings suggest that detailed palynological work in core could be used, alongside pressure data and seismic, to contribute to understanding of subsurface reservoir heterogeneity and ‘plumbing’. The reasons for palynological variations across the floodplain need to be investigated, but probably represent different abilities of parts of the floodplain to represent the hinterland and local floodplain floral communities.
This work will be the subject of a paper when all data has been collected and analysed.
References
Kerp, H., Blomenkemper, P. ., Abu Hamad, A., & Bomfleur, B. (2021). The fossil flora of the Dead Sea region, Jordan – A late Permian Garden of Delights. Journal of Palaeosciences, 70((1-2), 135–158. https://doi.org/10.54991/jop.2021.12
Stephenson, M H 2013. Returning Carbon to Nature; coal, carbon capture, and storage. Elsevier, August 2013, paperback ISBN: 9780124076716; eBook ISBN: 9780124076563.
Stephenson, M H, and Powell, J H. 2013. Palynology and alluvial architecture in the Permian Umm Irna Formation, Dead Sea, Jordan. GeoArabia, 18, 3, 17-60.
Stephenson, M H, and Powell, J H. 2014. Selected spores and pollen from the Permian Umm Irna Formation, Jordan, and their stratigraphic utility in the Middle East and North Africa. Rivista Italiana di Paleontologia e Stratigrafia, 120, 145-156.