. 2018 Jun 12;9:2201. doi: 10.1038/s41467-018-04423-1

Table 2.

Time averaged leakage rates of the three scenarios

Scenario	Time (year)	Time averaged leakage rate (% per year)^a
Scenario	Time (year)	P95	P50	P05
Offshore Well-Regulated	1	0.0005	0.0008	0.001
	100	0.0002	0.0004	0.0009
	1000	0.0001	0.0002	0.0006
	10,000	0.00005	0.00019	0.00063
Onshore Well-Regulated	1	0.001	0.002	0.005
	100	0.0007	0.002	0.004
	1000	0.0002	0.0009	0.003
	10,000	0.0002	0.0008	0.003
Onshore Poorly-Regulated	1	0.05^b	0.2^b	0.5^b
	100	0.02^b	0.06^b	0.2^b
	1000	0.002	0.008	0.02^b
	10,000	0.0007	0.002	0.003

The leakage values are expressed as a percentage of originally injected CO₂. Example times are presented at t = 1, 100, 1000 and 10,000 years

^aTime averaged leak rates are calculated by dividing the total cumulative leakage computed for the selected model time by the same number of model years. This results in an artificial linear rate of leakage which is constant from the start of injection to the selected time, and obscures the true variation in leakage rates over time (cf. Fig. 4)

^bResults that do not meet the 0.01% per year acceptability level^13,15. Notably, even on this simple metric, all well-regulated regions pass the simple acceptability test at all timescales. For the worst-case Poorly-Regulated Onshore Scenario, time-averaged leak rates are unacceptable in the short term, but at least 95% of the realisations give acceptable time-averaged leak rates over long time scales (several 1000 years)