1   Introduction

The public is progressively expressing greater alarm over greenhouse gas emissions, and more countries are taking steps to solve the problems brought about by climate change (IPCC, 2023; Breton and Sbragia, 2017; Basseches et al., 2022; Wei et al., 2021). China has committed to achieving a CO₂ emission peak before 2030 and carbon neutrality before 2060 (Xi, 2020). Carbon dioxide capture and storage (CCS) is a key technology for CO₂ emission reduction. CCS primarily means the injection of the captured high-purity CO₂ into selected safe geological structures for permanent storage. Generally, depleted oil and gas reservoirs, unminable coal beds, and deep saline aquifers are all possible formations for CO₂ storage (Zhang et al., 2014; Yang et al., 2017; Wang et al., 2023; Wong et al., 2007; Medici et al., 2019a, 2019b). Deep saline aquifers are widely distributed and are the most feasible alternative for CCS (Bachu and Adams, 2003; Li et al., 2023). Previous research showed that over 1 000 billion tons of CO₂ can be stored in deep saline aquifers of China (Dahowski et al., 2009; Guo et al., 2014). Over the decades, several pilot or commercial saline aquifer CO₂ storage projects have been ongoing or suggested; examples include the Sleipner and Northern Lights in the North Sea (Furre et al., 2017; Zhou et al., 2023; Mi, 2023), In Salah in Algeria (Ringrose et al., 2013), and Gorgon in Australia (Li et al., 2019). In China, the storage of CO₂ in saline aquifers remains in the exploratory stage. The Shenhua Ordos CCS project (Wu, 2013) and CNOOC's CCS project at the Enping 15-1 oilfield (Yi et al., 2023; Mi, 2023) are presently the only two demonstration projects that have been implemented, onshore and offshore, respectively.

Careful site selection, characterization, and injection practices are required for long-term security. After injection into saline aquifers, a combination of physical and geochemical processes occurs (IPCC, 2005; Seyyedi et al., 2016). The CO₂ plume migrates longitudinally and laterally within the reservoir. Comprehending the lateral migration of the CO₂ plume can facilitate answering the question of whether CO₂ can reach possible risk points. Several researchers have forecast the CO₂ migration range according to demonstration projects (Shi et al., 2019; Ahmadinia et al., 2020; Strandli et al., 2014; Liu et al., 2014). For example, Karstens et al. (2017) estimated the evolution of the CO₂ plume in the Sleipner Project through iterative history matching. Their long-term simulations showed that under the current injection scheme, CO₂ cannot reach the possible leakage pathway, namely, a chimney structure 7 000 m away from the injection well. Xie et al. (2015a, 2015b) forecast the dynamics of the plume in the Shenhua CCS Project by using numerical simulation. The maximum lateral distance of the CO₂ plume would reach approximately 1 000 m 100 years after the injection. Ziemkiewicz et al. (2016) used CO2-PENS, a system analysis tool, to assess the CO₂ plume radius in the Guantao and Donying formations of the Bohai Bay Basin. After ten years, the mean radius of the injection at 0.1 and 1 million tons per year (Mt/a) were 1.22 and 3.85 km, respectively.

Moreover, the effects of several parameters on CO₂ migration have been studied through analytical and numerical modeling (Birkholzer et al., 2015; Dai et al., 2018). Jing (2021) simulated the effect of formation slope and brine salinity using an improved simulator based on TOUGH2-ECO2N. In situations of high brine salinity and large formation slope, the CO₂ plume moves a longer distance. Sohal et al. (2021) analyzed the effect of geological parameters based on the characterization of the Hontomín saline aquifer. Heterogeneity in the permeability and porosity of the matrix slow down the migration of CO₂. Afanasyev et al. (2023) focused on the migration in a sloping aquifer during the post-injection period. They found that the shape of the plume at the time the injection well is shut in influences the subsequent migration. They also proposed a new simple relationship to estimate the maximum migration distance of the plume.

The study area, located in the largest Meso – Cenozoic superimposed basin in offshore China, is one of the most promising sites for CO₂ storage. However, few scholars have studied the migration of CO₂ in deep saline aquifers in this area. This paper aims to obtain the migration range of CO₂ under the geological conditions of the area. The goals are to utilize numerical simulation techniques to 1) describe the migration behavior of CO₂ in the aquifers during and after injection, 2) determine the effects of the geological and injection parameters on the CO₂ migration range, and 3) summarize the CO₂ migration range under definite reservoir and injection parameter changes. The migration range of CO₂ in deep saline aquifers can be considered the minimum safety distance between the injection and risk points when designing CCS projects in the study area.

2   Geological settings

The study area is situated in the greatest offshore sedimentary basin of China. The study area has tectonically undergone the evolution of a rifting stage, a thermal depression stage, and a regional subsidence stage. Studies reveal that the thickness of the Meso – Cenozoic stratum almost reaches 10 000 m. Saline aquifers are widely developed in the Miocene Series formation. According to the United States Department of Energy and the China Geological Survey, the CO₂ storage capacity in the deep saline aquifers of the basin is approximately over 100 billion tons of CO₂ (Dahowski et al., 2009; Guo et al., 2014). Hence, the basin is one of the most promising sites for storing CO₂ emissions from the southeastern coastal areas of China.

The study area mainly features large anticlines with a formation dip of less than 10°. The study area has two sets of high-quality regional seals and several sets of local seals. Potential saline aquifers are situated at depths of about 1 000 m and contain high-salinity brine (20 000 mg/L). The study area develops the braided river delta-front sub-facies. The lithology is primarily sandstone dominated by quartz. Large thicknesses and favorable properties also render it appropriate for CO₂ storage. The formation thickness, which is generally around 100 m, ranges from 10 m to 200 m. The porosity is about 20%, whereas the permeability varies greatly, ranging from 10 mD to 1 000 mD. Moreover, CO₂ can arrive at the supercritical state under the formation temperature and pressure conditions, with a temperature range of 58–63 ℃ and a pressure coefficient of 0.99.

3   Model establishment

3.1   Base model description

The CO2STORE module of the Schlumberger Eclipse 300 simulator is used in this study. The reactions between CO₂ and soil minerals are excluded from this study because mineral trapping takes a very long time, for example, hundreds or even thousands of years (IPCC, 2005). Three primary mechanisms, namely, structural trapping, residual trapping, and solubility trapping, are considered on the simulation time scale. The drainage and imbibition curves are computed using Killough's method (Killough, 1976) to represent residual trapping. The procedures proposed by Spycher and Pruess (2005, 2009) are used to determine the mutual solubilities of CO₂ and H₂O.

To study the migration behavior of CO₂ in the aquifers during and after injection, a homogeneous saline aquifer model is established based on the geological characteristics of the target CO₂ storage aquifers in the study area. The aquifer parameters of the base model are presented in Table 1, and the relative permeability curves used for simulation are shown in Figure 1. The model has dimensions of 5 000 m×5 000 m×100 m and an injection well at the center, as illustrated in Figure 2. CO₂ is continuously injected into the aquifer for 20 years, and then the injection well is shut in while the simulation proceeds for another 100 years. In the base simulation scenario, a vertical well with a injection rate of 0.2 Mt/a of CO₂ is implemented in this study.

Figure  1  Relative permeability curves of CO₂ and brine phase

Download: Full-Size Img

Figure  2  Mechanism model of CO₂ storage in a saline aquifer

Download: Full-Size Img

3.2   Case studies

Geological data from the seismic survey and petroleum exploration and production wells in the promising CO₂ storage area indicate that reservoir thickness, permeability, and formation dip vary within a certain range. Some uncertainty regarding the brine salinity of the aquifers may exist owing to the lack of specialized CCS pilot wells. Thus, four geological parameters are considered in the case studies: reservoir permeability, reservoir slope angle, reservoir thickness, and formation brine salinity. In addition, the influence of injection rate and horizontal well length are studied. The changes in the geological and injection parameters are presented in Table 2, and the sensitivities of each parameter to the migration range are studied.

4   Results

4.1   CO₂ migration in the base model

Figures 3 and 4 show the CO₂ mole saturation profiles of the base case at four simulation time steps: 1, 10, 20, and 120 years. The injected CO₂ is driven upward due to the buoyancy effect from the density difference with the formation brine. After reaching the impermeable overlaying cap layers, the CO₂ begins to migrate laterally. Some of the CO₂ dissolves into the aquifer as the plume moves longitudinally and laterally. Residual trapping happens when saturation decreases due to the buoyancy effects and dissolution processes. Consequently, the CO₂ phase saturation gradually decreases from the injection well outward.

Figure  3  CO₂ distributions in the base simulation case

Download: Full-Size Img

Figure  4  Plane migration of CO₂ in the base simulation case

Download: Full-Size Img

For convenience, the overall lateral distance of the CO₂ plume is defined as the migration range, as shown in Figures 3 and 4. The migration range of the CO₂ varies over simulation time, as shown in Figure 5. Results indicate that the CO₂ plume primarily migrates during the injection period rather than the post-injection period, although the duration of the post-injection period is five times longer than that of the injection period. Specifically, the migration range at the end of the 20-year injection period accounts for over 70% of the total range within the simulated time scale.

Figure  5  Variation of CO₂ migration range with simulation time

Download: Full-Size Img

4.2   Effect of geological parameters

4.2.1   Reservoir permeability

Figure 6 presents the distribution of CO₂ mole fraction in aquifers under diverse permeability conditions after a 120-year simulation. High aquifer permeability increases the potential migration range of the CO₂ plume. The longitudinal and lateral migration both become easier with an increase in aquifer permeability. Meanwhile, more CO₂ dissolves in the brine of the upper formation, leading to convective mixing. In high-permeability aquifers, dense brine dissolved with CO₂ falls, creating plumes that resemble fingers. The fingers become more evident as the permeability increases, which may slow down the lateral migration of CO₂.

Figure  6  Distributions of CO₂ in saline aquifers with various permeabilities (t = 120 a)

Download: Full-Size Img

Figure 7 shows the migration range of the CO₂ plume changes similarly with permeability at different simulation time steps. When the permeability is less than 200 mD, the migration range substantially increases with the increase of permeability. As the permeability increases further, the increase of the migration range becomes less substantial due to the influence of convective mixing. Moreover, the CO₂ plume migration range increases substantially during the injection period, while the expansion of the CO₂ plume is limited after the injection well is shut in. Taking the case with a 200 mD permeability as an example, the migration ranges at 10, 20, and 120 years are 1 550, 1 850, and 2 350 m, respectively. The migration range during the injection period accounts for 79%.

Figure  7  Variation of CO₂ migration range with reservoir permeability

Download: Full-Size Img

4.2.2   Dip angle

The stratum of the target CO₂ storage aquifers is almost horizontal, with a slight dip in areas. The distributions of the CO₂ mole fraction in aquifers dipping at 0° – 10° are presented in Figures 8 and 9. The vertical scale of the map is 1:3.

Figure  8  Distributions of CO₂ in saline aquifers with various dip angles (t = 120 a)

Download: Full-Size Img

Figure  9  Variation of CO₂ migration range with dip angle

Download: Full-Size Img

Injection into dipping saline aquifers causes CO₂ to behave differently compared with that in horizontal aquifers. When CO₂ arrives at the bottom of the overlaying cap layer, it proceeds to move upward owing to the influence of buoyancy. Hence, more of the injected CO₂ migrates to the updip direction. During the process, more fresh water comes into contact with the CO₂ plume, which can enhance the process of dissolution.

The simulation presents that the greater the dip angle of the formation, the larger the upward migration range, while the smaller the downward migration range. However, the effect of reservoir dip angle on the entire range of CO₂ plume migration is not substantial, as shown in Figure 8. After the 120-year simulation, the migration range of CO₂ for the horizontal case is 2 200 m, whereas that in the case of an aquifer with a dip angle of 10° is 2 488 m, which is a 13% increase. Thus, aquifers with a slight dip angle, such as 1°–5°, can also be appropriate for CO₂ storage because the effect on the entire migration range is restricted.

4.2.3   Reservoir thickness

The injected CO₂ first migrates longitudinally due to the influence of buoyancy and then laterally after arriving at the bottom of the overlaying cap rocks. The thickness of the reservoir influences the longitudinal migration of CO₂, which may also have a bearing on lateral migration.

At the end of the simulation (t =120 a), the CO₂ plume distributions in saline aquifers with diverse reservoir thicknesses range from 20 m to 200 m, as shown in Figure 10. The results reveal that when the reservoir thickness is less than 120 m, the CO₂ migration range continuously increases with the increase of thickness. When the reservoir thickness increases to 120 m, the injected CO₂ is primarily distributed at the top and middle of the reservoir despite the same injection amount. The thickness of the reservoir no longer influences the lateral migration. The maximum migration range of the CO₂ plume is approximately 2 250 m.

Figure  10  Distributions of CO₂ in saline aquifers with various reservoir thicknesses (t = 120 a)

Download: Full-Size Img

Figure 11 shows that the 20-year injection period remains the primary period for migration, accounting for 78%–90% of the total. The CO₂ lateral migration range varies with reservoir thickness over diverse observation times. In the early stages of injection, reaching the bottom of the cap rocks takes a longer time in thick reservoirs, causing smaller lateral migration ranges. The plume migration range in the 200 m reservoir is 15% smaller than that in the 50 m reservoir after CO₂ injection for 1 year. In the middle to late stages of the simulation, the lateral migration range of CO₂ is larger in thick reservoirs, but its effect is not substantial.

Figure  11  Variation of CO₂ migration range with reservoir thickness

Download: Full-Size Img

4.2.4   Brine salinity

Theoretically, low brine salinity is beneficial to CO₂ dissolution. The migration of CO₂ in aquifers with salinity ranging from 0 mg/L to 50 000 mg/L is simulated.

The effect of brine salinity on migration range is not substantial, as shown in Figures 12 and 13. At the end of the simulation, the range of the high-salinity case is slightly larger than that of the low-salinity cases. The primary reason is that the solubility of CO₂ decreases in high-salinity brine, causing more CO₂ to exist as a free phase. The CO₂ plume migrates farther over time, and the migration range during the injection period accounting for 77%–90% of the total.

Figure  12  Distributions of CO₂ in saline aquifers with various brine salinities (t = 120 a)

Download: Full-Size Img

Figure  13  Variation of CO₂ migration range with brine salinity

Download: Full-Size Img

4.3   Effect of injection parameters

4.3.1   Injection rate

Based on worldwide pilot and commercial saline aquifer storage projects, the yearly CO₂ injection rate of a single well ranges from 0.1 Mt to 1 Mt. Five different injection rates within this range are selected for sensitivity study in this section. A tolerant pressure limit is implemented in this study. When the average reservoir pressure is lower than the fracture pressure of the cap rock, CO₂ can be continuously injected. None of the simulated scenarios reach the pressure limit during the injection.

Figure 14 presents the distributions of CO₂ after a 120-year simulation. A remarkable difference is noted in the shape of these cases. At high injection rates, CO₂ is more likely to displace the brine in a piston-like mode rather than form a plume. Therefore, the CO₂ in the middle and lower parts of the reservoir migrates farther, whereas the overall migration range along the top of the reservoir only slightly increases. The injection rate is more influential on migration at low levels, as shown in Figure 15. The migration range at the end of the simulation increases by 26% when the annual injection rate of CO₂ increases from 0.1 Mt to 0.5 Mt. For higher injection rates, the migration range remains basically unchanged. The higher average saturation and pressure in scenarios with a higher injection rate may help CO₂ migrate further over the next few hundred years. Overall, the increase in migration range during the post-injection period is restricted, and the CO₂ migration range remains within 3 000 m.

Figure  14  Distributions of CO₂ in the saline aquifer at various injection rates (t = 120 a)

Download: Full-Size Img

Figure  15  Variation of CO₂ migration range with injection rate

Download: Full-Size Img

4.3.2   Horizontal well length

The effect of injection well type, including vertical and horizontal wells, is discussed in this section.

Compared with a vertical well, the CO₂ plume migrates farther when injected with a horizontal well, as illustrated in Figures 16 and 17. The migration range increases with the increase of horizontal well length. In particular, the CO₂ plume is no longer symmetrically distributed in horizontal injection cases. A higher injection pressure difference at the heel of the horizontal well may cause an increase in CO₂ saturation and migration range. For a horizontal well of 2 000 m, the migration range is almost 2 850 m at the end of the simulation, which represents a 33% increase compared with a vertical well.

Figure  16  Distributions of CO₂ in saline aquifers under various horizontal well lengths (t = 120 a)

Download: Full-Size Img

Figure  17  Variation of CO₂ migration range with horizontal well length

Download: Full-Size Img

Based on various geological simulation cases in Section 4.2, CO₂ migration range and permeability, dip angle, thickness, and salinity always present a certain degree of positive correlation. To clarify the substantial parameter, the relative increases in the migration range under the variation of each parameter are summarized in Table 3. Within the statistical range of the above parameters, permeability has the most substantial effect, whereas the influences of other parameters are within 10%. Then, the CO₂ migration range can be estimated considering the comprehensive effect of various geological parameters. In extreme cases, the migration range of CO₂ may be 45% larger or 51% smaller than that in the base case. Therefore, a safety distance of 1 053–3 117 m from risk points during site selection in the study area is recommended. For safety reasons, a distance of over 3 000 m is more preferable.

Diverse injection scenarios have also been analyzed in Section 4.3. Long horizontal wells may also lead to greater risks, although they can increase the injectivity of CO₂. In this way, the location of the injection well should be further optimized within the selected site after determining the horizontal length based on the injection amount requirement. For example, a 33% greater distance should be kept for a 2 000 m horizontal well.

5   Conclusions

Simulations of CO₂ storage in saline aquifers reveal that CO₂ mainly migrates during the injection period rather than the post-injection period, and the migration range of the injection period accounts for more than 70% of the total. Based on the geological characteristics of the saline aquifers in an offshore sedimentary basin, the overall range of CO₂ migration in saline aquifers is approximately within 3 000 m, considering various geological and injection conditions.

The effect on the migration range follows the order permeability > formation dip angle > reservoir thickness, whereas brine salinity has a slight effect. As reservoir permeability, dip angle, and thickness increase, the migration range generally exhibits an increasing trend and stabilizes when the above factors increase to a specific value.

Injection rate and well type substantially affect the migration of CO₂. When CO₂ is injected at a low rate, the migration range increases with the injection rate increasing. At high rates, the additional increase in migration range is restricted. However, the distribution of CO₂ varies from that at low rates, making it more likely to displace the brine in a piston-like mode. As for horizontal well injection, the CO₂ plume is no longer symmetrically distributed due to the pressure difference between the toe and the heel. The longer the horizontal well length, the farther the CO₂ migration.