Comment on "fate of rising CO2 droplets in seawater".

I welcome the comment by Alendal et al. (1), and the opportunity to clarify some issues. My paper (2) focuses on theoretical calculation of the dissolution rate of a single rising CO2 droplet (or noninteracting droplets) in the context of carbon sequestration in oceans. The calculated dissolution rate agrees with the in situ experimental data of Brewer et al. (3). The comment (1) does not raise any issue on the fundamental theoretical treatment. My paper (2) also treated a few other issues but in less detail. Alendal et al. (1) raise four points, mostly on these other issues, or on issues not treated by my paper (2), such as biological impact. The points are discussed below in sequence. 1. The possibility for injected CO2 to induce an eruption similar to lake eruptions. Because rising CO2 droplets injected into ocean water would convert to gas bubbles at shallow depth (about 400 m but depending on temperature), I (2) suggested that to “avoid the formation of rising CO2 droplet plumes and to avoid CO2-driven eruptions, injection of CO2 liquid into the shallow oceans needs to be controlled, using low injection rates and dispersing CO2 liquid to a large volume of seawater. Safer injection schemes include injection of CO2 liquid to great enough depths that liquid CO2 density is greater than ambient seawater so that CO2 liquid would sink and gradually dissolve in seawater.” Alendal et al. (1) quoted the IPCC report (4) that “there in no known mechanism that could produce an unstable volume of water containing 2 MtCO2 at depth shallower than 500 m, and thus no mechanism known by which ocean storage could produce a disaster like that at Lake Nyos”. The IPCC report (4) is an authoritative document on carbon storage. Nonetheless, details of the IPCC report are not necessarily accurate and quoting the IPCC report does not necessarily end a debate. For example, the above quotation of the IPCC report cites two million tons of CO2 as the amount released in the 1986 eruption of Lake Nyos, likely based on Kling et al. (5) (although another reference was cited, which does not contain the information and might be a typographical error). However, the issue was revisited after the work of Kling et al. (5) and consensus was reached that the amount of CO2 erupted was about 15% of the above quoted value (6). Furthermore, such an amount of CO2 is not necessary for a CO2-driven eruption; the amount of CO2 released by the 1984 Lake Monoun eruption was much less. One may envision both types of scenarios: (i) slow injection of individual droplets would not induce an eruption, and (ii) as the IPCC report pointed out “if somehow large volumes of liquid CO2 were suddenly transported above the liquid-gas-phase boundary, there is a possibility of a selfaccelerating regime of fluid motion that could lead to rapid degassing at the surface” (4, p 308). The possibility and dynamics of ocean eruptions powered by the sudden formation of many bubbles have been discussed (7, 8). The threshold between the two extremes has not been mapped out. Contrary to the claim of Alendal et al. (1), at depth shallower than 2000 m, the plume water would be less dense if CO2 is in the form of droplets or bubbles instead of the dissolved form. I maintain that (i) new research is necessary for understanding the condition under which safe injection can be achieved, and (ii) it is necessary to exercise caution in injecting a large amount of CO2 into the deep ocean. As long as the injection design is carefully evaluated, escape of CO2 to the surface through bubble plumes is avoidable. Claiming that such caution is unnecessary would not be wise.