CO2 storage monitoring · seafloor gravimetry · Sleipner-class survey design
CLAIMABLE — four core gates pass, and recovery verified against a known truth The numbers below for this monitoring programme are supported by the data.
Gates
Check
Result
PASS
licensing
0 (< 95.0 pct) observation must be plausible under the prior
0.001302 (0.001 < pp < 0.009) the only gate that tests reality
PASS
recovery
1 (cover >= 1.00 and worst miss <= 3 sigma) worst 0.3 sigma — the gate the other four cannot be
What the data supports
CO2 accounted for inside the complex
7.4 +/- 7.1 Mt (declared injected: 10.0 Mt)
inventory reconciles?
YES — the declared inventory sits inside the 95% interval
CONTAINMENT — smallest leak this survey can detect
5.21 Mt in a 400 m body 2 km out — but only if that body is at least 72% saturated
so an emptier leak of that size is invisible
a 400 m body below 72% saturation cannot be ruled out at all, and no containment claim covers it
the recovered picture, judged honestly
the posterior mean exceeds the physical saturation ceiling in 0% of cells — a Gaussian prior is unbounded, so the MAP picture is not physical. The mass limits are integrals and stay valid; being unbounded makes them conservative, never optimistic
the honest limit of gravity ALONE
per-column sigma is ~0.9 Mt against a peak column of 0.15 Mt — this survey constrains INTEGRALS, not maps. That is why Sleipner runs gravity alongside seismic, and any claim to map a plume from 30 benchmarks alone would be the prior talking
what a whole-area bound would have said
< 106 Mt outside — REFUSED, it is 1065% of the injected mass and is set by the prior, not the data
inventory +/-7.1 -> +/-6.2 Mt; 400 m leak threshold 5.21 -> 4.90 Mt (6% better for a 2.7x quieter instrument — this survey is limited by its geometry, not its noise)
The survey being judged
benchmarks
30 on the seafloor, 500 m spacing
repeatability
3.0 uGal (Sleipner 2002-2009 epochs)
reservoir
950-1050 m depth, porosity 0.36
in-situ densities
CO2 700 vs brine 1050 kg/m3 -> bulk contrast -0.126 g/cc when saturated
peak signal
19.0 uGal against 3.0 uGal noise
The 3.0 -> 1.1 uGal row above is computed, not assumed: that is the trade a monitoring budget turns on, and it is knowable before a single benchmark is deployed.
Leak-detection capability, by body size (2 km outside the complex)
A containment claim is only honest for bodies in the VISIBLE rows. For the rest the survey is silent, and saying otherwise would be reporting the prior as if it were a measurement.
What this analysis is, precisely
survey parameters
published Sleipner monitoring programme
observations
SIMULATED from the declared plume above
Sleipner measured data
NOT used — licensed per user and not sellable; deliberately excluded
what the numbers mean
what this survey DESIGN can prove
A capability analysis is what an operator needs before committing to a monitoring programme and what a regulator needs to judge one. Substituting a licensed real dataset changes the inputs, not the method.
Where the model is real
White dashed circle: the licensed storage complex. Black dots: seafloor benchmarks. Panels 1 and 2 share a colour scale on purpose: 30 stations against 4,500 unknowns cannot resolve a picture, and the recovered mean is largely ringing. That is the honest state of the art, and it is why the deliverable is panel 4 — the most CO2 that could be hiding in each column without this survey seeing it — and not the pretty map.
