Scholarly article on topic 'Numerical study on the rise of pCO2 in seawater by the leakage of CO2 purposefully stored under the seabed'

Numerical study on the rise of pCO2 in seawater by the leakage of CO2 purposefully stored under the seabed Academic research paper on "Earth and related environmental sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Energy Procedia
Keywords
{"Subsea geological storage of CO2 " / "Leakage of CO2 " / "Two-phase flow" / "Dissolution of CO2 " / "Bubbles and droplets"}

Abstract of research paper on Earth and related environmental sciences, author of scientific article — Yuki Kano, Toru Sato

Abstract In this study, numerical simulations we re conducted to investigate the rise of pCO2 in the ocean on a continental shelf by the leakage of CO2, which is originally stored in the aquifer under the seabed, in an extreme case. We conducted parameter studies in simple rectangular solid domains with uniform background flows to see the impacts of conditions of CO2 leakage and those of seawater. The results show that temperature and the size of bubbles and droplets have reasonable impacts on the change of TCO2. Moreover, the background TCO2 makes a larg e difference in the rise of pCO2.

Academic research paper on topic "Numerical study on the rise of pCO2 in seawater by the leakage of CO2 purposefully stored under the seabed"

MUhonh riwuNudMHtiHUiiiii

^"SdenceDlrad:

Energy Procedía 1 (2009) 1909-1916

Energy Procedía

www.elsevier.com/locate/procedia

GHGT-9

Numerical study on the rise of pCO2 in seawater by the leakage of CO 2 purposefully stored under the seabed

Yuki KANOab, Toru SATOa*

a Department of Ocean Technology, Policy, and Environment, University of Tokyo, 5-1-5 Kashiwa -no-ha, Kashiwa 277-8563, JAPAN bCurrent Adress: National Institute of Advanced Industrial Science and Technology, 1-1-1 Higashi, Tsukuba 305-8567, JAPAN

Abstract

In this study, numerical simulations were conducted to investigate the rise of pCO2 in the ocean on a continental shelf by the leakage of CO2, which is originally stored in the aquifer under the seabed, in an extreme case. We conducted parameter studies in simple rectangular solid domains with uniform background flows to see the impacts of conditions of CO2 leakage and those of seawater. The results show that temperature and the size of bubbles and droplets have reasonable impacts on the change of TCO 2. Moreover, the background TCO2 makes a large difference in the rise of pCO2. © 2009 Elsevier Ltd. All rights reserved.

Keywords: Subsea geological storage of CO 2; Leakage of CO2; Two-phase flow; Dissolution of CO2; Bubbles and droplets

1. Introduction

Geological storage of CO2 under the seabed is one of the methods to mitigate the global warming. This method is considered to reduce CO2 in the atmosphere and to attenuate the acidification in the surface water. However, this method has a risk of CO2 leakage, which may cause CO2 dissolution in the seawater resulting in the negative impacts on the organism in the ocean near the leakage site. On the other hand, if most of leaked CO2 comes back to the atmosphere, the mitigation effect of global warming would be compromised. Therefore, it is important to know the behaviour of leaked CO2 and its biological impacts in the ocean after leaked from the seabed. In this study, numerical simulations were conducted to predict the rise of partial pressure of CO2 (ApCO2) in the ocean continental shelf by the leakage of CO2, which is originally stored in the aquifer under the seabed, in a very extreme case, such as a large fault connects the CO2 reservoir and the seabed accidentally by big earthquakes or other large diastrophisms [1] . The behavior and dissolution of CO2, which take the forms of bubbles or droplets depending on the pressure and the temperature of the seawater at depth, were numerically simulated during the rise in seawater flows. Also simulated were the advection-diffusion of dissolved CO2 and the chemical equilibrium in the seawater related to ApCO2. The two-phase flow simulations by a Full-3D model were conducted with simple rectangular solid

* Corresponding author. Tel.: +81-4-7136-4727; fax: +81-4-7136-4727.

E-mail address: sato-t@k.u-tokyo.ac.jp .

doi:10.1016/j.egypro.2009.01.249

water columns with uniform background flows for the parameter studies to study the impacts o f the conditions of CO2 leakage and those of seawater on the rise of the dissolved CO2 concentration (ATCO2) and ApCO2.

2. Methods

A numerical model necessary for this study is similar to those developed to predict the behaviour of droplets of CO2, which is purposefully injected in the deep ocean, such as Sato and Sat o [2]. Their method adopted an Eulerian -Lagrangian two -phase model, that is, it used a finite difference method for the continuous seawater phase, and the movement of the dispersed phase was analyzed by solving the motion equation of an individual bubble or droplet.

Because the method of Sato and Sato [2] was developed to simulate the movement and dissolution of CO2 droplets only, it is necessary to change the drag and the mass trans fer coefficients when dealing with bubbles in more shallow water. When a bubble rises in water, the water depth, or in other words, the hydrostatic pressure and the temperature, controls its volume and shape; a sphere, an ellipsoid, or a spherical cap. The method was modified to adopt those shape change effect. The target space in this study was limited to the ocean above the seabed, and did not include subsea underground. Leakage rate was based on the simulations of the CO2 flow in geological formations found by RITE [1].

For the biological impacts of CO2 in the ocean, we focused on ApCO2 following Kikkawa et al. [3] who elucidated that the biological impacts of CO2 in the ocean does not refer to pH or TCO2 but to pCO2. Kita and Watanabe [4] tentatively proposed that the predicted non-effective concentration (PNEC) of CO2 is 500 ppm (corresponding to ^-atm) in ApCO2 from the background value. I f ApCO2 is lower than this, it is believed that there is no biological impact for marine organisms in the deep ocean. For calculating pCO2 by using TCO 2, temperature, pressure, and salinity, the method developed by Carbon Dioxide Info rmation Analysis Center [5] was used, with a set of coefficients obtained by Roy et al. [6].

3. Computational Condition s

Figure 1 shows TCO2-pCO2 curves, where lines are correspondent to various sets of temperatures and salinities. The curves indicate that even though ATCO2 is the same, larger those background values leads to larger ApCO2, especially with larger background temperature and TCO2. In our former simulations [7], those conditions of seawater were constant in the all cases. Moreover, initial TCO2 was set to be as large as that of the deep ocean, although their target was the rather shallower sea on a continental shelf, in which, TCO2 is generally much lower. In this study, a more realistic vertical profile of TCO2 around Japanese coast was chosen as the background condition, as well as the one used in the former simulations [7] for a comparison.

Figure 2 shows the schematic view of the simulation for the parameter studies. The computational domain was rectangular solid and ranges from a sea surface to a flat seabed in the vertical direction. It was assumed that the uniform flow comes in at the left -hand -side boundary, and the flow was maintained by setting it constant as a Dirichlet boundary condition. At the sea surface and seabed, free-slip was assumed. The CO2 leakage takes place on a fault band, which is perpendicular to the uniform inflow, and 25 m wide in the flow direction and infinity long in the direction of periodical boundary conditions.

Table 1 shows the conditions of the parameter studies. These simulations were conducted to investigate the impacts of leakage depth, size of bubbles/droplets, current velocity, and the background profiles of temperature, salinity, and TCO2, on the ATCO2 and ApCO2 from the background values.

The 100 percent CO2 leaks in the form of bubbles at the depth of 200 m, and of droplets at the depth of 500 m. Initial sizes of leaked bubbles and droplets follow the Gaussian distributions of which the averages || take values in Table 1 and the variances o=0.21. The diameter in Case 11 is the one for droplets, the mass of which is equal to that of bubbles in Case 1 .

Figure 3 shows the vertical profile of temperature, salinity, and TCO2 cited in Table 1 and the surface concentration of CO2, that is, the gas solubility in the bubble cases and the equilibrium concentration of CO2 coexisting with CO2 hydrate in the droplet cases, with respect to the water depth at the various temperature profiles. The profiles, T SL, TDL, S SL, SDL, CSL, and CDH, are typical near the Japan ese coast. C SL and CDH were provided by

Ono [8], who measured them in the observation voyag e mentioned in Ono et al. [9]. The other profiles were the parallel s hifts of the typical ones to study their own effects on CO2 dissolution and chemical equilibrium in the seawater. CSX is rather unreal for the water on continental shelves; i.e. TCO2 is as high as that in the deep ocean and this was used by Kano et al. [7]. The tendency of the solubility dramatically changes at large depths, because hydrate film covers the CO 2 droplets .

Figure 1 Correlation between TCO 2 and pCO 2 for various pairs of temperature in oC and salinity in %o; a: (2, 34), b: (10, 34), c: (22, 34), d: (10, 32), and e: (10, 36)

Periodical Boundary

25m / N /

25m / /

I \ °o

Uniform \ t>o o

Flow / 1 o o 0 CO£ Hume o °

200 or

500m ■ ■ / 25m /

Figure 2 Schematic view of com putational domain

Table 1 Calculation conditions. The profile indicators refer to those i n Figure

Case Seabed Inflow Mean initial Temperature Salinity TCO2

depth (m) velocity (m/s) diameter (x 10" m) profile profile profile

1 200 0.05 2.0 TSL SSL CSL

2 200 0.10 2.0 TSL SSL CSL

3 200 0.05 4.0 TSL SSL CSL

4 200 0.05 1.1 TSL SSL CSL

5 200 0.05 2.0 TSH SSL CSL

6 200 0.05 2.0 TSL SSH CSL

7 200 0.05 2.0 TSL SSL CSH

8 200 0.05 2.0 TSL SSL CSX

9 500 0.05 2.0 Tn. SDJ. Cn.

10 500 0.05 1.1 TDL Srj. CDL

11 500 0.05 0.36 TDL Srj. CDL

12 500 0.05 2.0 TDH SDJ. Cn.

13 500 0.05 2.0 TDL SDH CDL

14 500 0.05 2.0 Tdl CDH

5 10 15 20 Temperature (°C)

Figure 3 Background vertical profiles of temperature (a), salinity (b), continued

0 0.02 0.04 0.06

Surface concentration of C02 (mass fraction)

Figure 3 Background vertical profiles of TCO2 (c), and the surface concentration of CO2 (d); a: Case 1, b: Case 5, c: Case 9 and d: Case 12

4. Results and Discussion s

Figure 4 shows the contour maps of volume fraction of undissolved CO2 after 10 hours of the start of the leakage. The plumes of CO2 bubbles and droplets show leanings, following the uniform flow and dissolving into seawater during their rise. In Case 9, which is the control condition of the droplet-cases, CO2 plume shows larger leaning angles because of smaller upward velocity of droplets than in Case 1, which i s the control condition of the bubble " cases. All CO2 dissolves within the vertical distance of about 100 m in both cases, as is the same in all other present cases, without coming back to the air. Therefore, it is indicate d that the ocean could be a buffer against the negative impacts on the effect of the restraint of the global warming or on direct human health, which might be caused by the leaked CO2.

Figure 5 shows the contour maps of ATCO2. Because of the infinity long leakage-source band and the uniform inflow, ATCO2 is almost equal in the downstream direction. These conditions also lead to larger ATCO2 in deeper layer. ATCO2 is larger in Case 9 than in Case 1 in the deepest layer, which may results from that droplets tend to be floating longer in the lower water column because of their smaller upward velocity.

Figures 6 and 7 show the vertical profiles of simulated ATCO2 and ofApCO2, respectively, in each case at x=750 m. The line of Case 1 almost overlay s with thos e of Cases 6, 7, and 8 in Figure 6(a), and so does the line of Case 9 with those of Cases 13 and 14 in Figure 6(b). Similarly, the line of Case 1 overlays with that of Case 6 in Figure 7(a), and so does the line of Case 9 with that of Case 13 in Figure 7(b).

Figure 6 shows that the salinity and the background TCO2 profiles are hardly meaningful in the change of TCO2 in the water, while the current velocity and the temperature profile have reasonable impacts on it. Whereas Case 5, in which temperature is higher than that in Case 1, shows slightly smaller ATCO2 than Case 1 in the deepest layer, Case 12, which is the high-temperature-condition case of droplets, shows larger ATCO2 than Case 9 in the deepest. This difference may be because of the dramatic change of tendency in the CO2 solubility, as is shown in Figure 3(d). In shallower depth, CO 2 solubility is smaller when temperature gets higher. However, in deeper depth, it shows the opposite tendency for the temperature to the gas solubility with the presence of the hydrate.

The impact of the size of bubbles or droplets is also noticeable. It shows large non-monotonous effects on ATCO2. Total CO2 dissolution from bubbles or droplets is determined by the balance of their total surface area and their rising velocity, which affects the mass transfer coefficient of them. For example, if initial size of bubbles is smaller, then the total surface area would be larger, although their rising velocity might be smaller and so might be the mass

transfer coefficient. Numerical simulation is an effective method to study which of these effects would be larger a size.

Figure 7 shows that although ApCO2 profiles appear to be almost in accordance with those of ATCO2 in Figure 6, some parameters make non-negligible differences on them. In particular, the background TCO2 makes a large difference in ApCO2, despite of its little impact on ATCO2. Case 5 and Case 12, which are the high-temperature -condition cases, also show larger ApCO2 than each standard case. These may result from the effect of TCO2 and temperature on pCO2, as is shown in Figure 1.

Furthermore, ApCO2 in the all cases were smaller than the present PNEC, 500 ppm, which indicates that the impact of the CO 2 leakage may be insignificant for marine organisms, such as zooplanktons, even in such an extreme case . However, t hese simulations leave some simplified assumption, such as rectangular solid domains, infinity long leak -source band, uniform background flows, and parameterizations for bubbles and droplets. These properties expect detailed studies for more realistic evaluation of the impacts of the leaked CO2. The site conditions such as temperature and background TCO2 also should be carefully examined for the accurate assessment with due regard for their large impacts, as are shown in this study.

-300 ?

s-'-500

0 200 400 600 800 1000 X(m)

VOID: 1.0E-09 6.3E-09 4.0E-08 2.5E-07 1.6E-06 1.0E-05

Figure 4 Contour maps of calculated volume fra ction of CO bubbles/droplets at 10 hours after the start of leakage in Case 1 (a) and Case 9 (b)

200 400 600 800 ^ooS00 X(m>

Figure 5 Contour maps of calculated ATCO2 in kg/m at 10 hours after the start of leakage in Case 1 (a) and Case 9 (b)

1-150 Q

-200 10

Rise of TC02 (kg/m )

Figure 6 Vertical profiles of calculated ATCO2 in kg/m at x = 750 m for bubble cases (a) and for droplet cases (b)

£-150 G

Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 Case 7 Case 8

(a) Rise of pC02 (ppm) (b)

Figure 7 Vertical profiles of ApCO2 in ppm at x = 750 m for bubble cases (a) and for droplet cases (b)

5. Conclusions

In this paper, we conducted numerical simulations by a Full-3D two-phase flow model on the rise of pCO2 in the shallow ocean caused by the leakage of CO 2. CO2, which is geologically sequestrated under the seabed earlier, seeped from the seabed [1] in the forms of bubbles or droplets depending on insitu pressure and temperature, and dissolves into the seawater during its rise. Dissolved CO2 was calculated by the advection-diffusion equation and followed the chemical equilibrium.

As a result, t he parameter studies show that the current velocity and the temperature profile have reasonable influences on the change of TCO 2 in the water, whereas the salinity profile is hardly meaningful. The size of bubbles/droplets is very important, because its effects on the rise of TCO2 are not monotonous. In addition, the background TCO2 makes a large difference in the rise of pCO2, although it has little impact on the change of TCO2.

It seems because of a nonlinear relation between pCO2 and TCO2. It is also found that all the CO2 dissolves within about 100 m, and does not come back to the air in this study. T his indicates that the ocean could be a buffer against the leaked CO2, which, otherwise, may compromises the effect of the mitigation of the global warming or may affects the human health directly. Moreover, the rises of pCO2 in the all cases were smaller than 500 ppm, which is the value of a tentatively proposed PNEC, under the present simulation conditions. Therefore, it was suggested by the model simulations that the impact of the leaked CO2 on marine organisms may be not significant, even under such an extreme case. It is to be noted, however, that the present simulations are rather conceptualistic, assuming simplified conditions, such as rectangular solid water columns, an infinity long leak-source band, uniform inflows and the parameterization s for bubbles and droplets. These conditions should be carefully examined to investigate more realistic impacts of the leakage of CO2, adding to the site conditions such as temperature and background TCO2, which showed their large effects on the ris e of pCO2 in this study.

More realistic simulation with the topography and the tidal current around Japanese coast is to be conducted by the authors.

6. References

1. Research Institute for Innovative Technology for the Earth (RITE), Annual Report of CO2 Geo logical Sequestration Project (2004), No. 2, 158-213 (in Japanese).

2. T. Sato and K. Sato, Numerical prediction of the dilution process and its biological impacts in CO 2 ocean sequestration, J. Marine Sci. Technol. 6 (2002), 169 -180.

3. T. Kikkawa, J. Kit a, and A. Ishimatsu, Comparison of the lethal effect of CO2 and acidification on red sea bream during the early developmental stages, Marine Pollution Bulletin 48 (2004), 108 -110.

4. J. Kita and Y. Watanabe, Impact assessment of high -CO2 environment on marine organisms, Proc. 8th Int. Conf. on Greenhouse Gas Control Technol. CD-ROM, 2006.

5. Carbon Dioxide Information Analysis Center, http://cdiac.ornl.gov/oceans/co2rprt.html.

6. R.N. Roy, L.N. Roy, K.M. Vogel, C. Porter-Moore, T. Pearson, C.E. Good, F.J. Millero, and D.M. Campbell, The dissociation constants of carbonic acid in seawater at salinities 5 to 45 and temperatures 0 to 45 deg C , Marine Chemistry 44 (1993), 249-267.

7. Y. Kano, T. Sato, J. Kita, Numerical Simulation on biological impact of leakage of CO 2 sequestrated under the seabed, Proc. 8th Int. Conf. on Greenhouse Gas Control Technol. CD-ROM, 2006.

8. T. Ono, Personal communication, 2007.

9. T. Ono, H. Kasai, T. Midorikawa, Y. Takatani, K. Saito, M. Ishii, Y.W. Watanabe, and K. Sasaki, Seasonal and interannual variation of DIC in the Oyashio mixed layer - a climatological view, Journal of Oceanography 61 (2005), 1075-1088.