Is fiber-optic VSP sensitive to cold-CO₂ Vp changes in shale-oil EOR?

Language

English version. 中文版 → 页岩油冷 CO₂ EOR：光纤 VSP 对时移 Vp 真的敏感吗？

The question. Shale-oil operators are running more and more CO₂ EOR (huff-n-puff and continuous). Intuition says: fiber-optic VSP (DAS-VSP) should be relatively sensitive to the time-lapse $V_P$ change after cold CO₂ injection. Is that right?

Verdict: half right, but oversimplified. Three sub-claims hold up; the rock-physics step is where the intuition leaks. And the exact triple — cold CO₂ × shale-oil EOR × DAS-VSP 4D $V_P$ — is almost a blank in the published literature.

The claim, decomposed

Sub-claim	Holds up?
① CO₂ / rich-gas EOR is a growing field practice in US tight oil (Bakken, Eagle Ford)	✅ Yes — 11 Bakken EOR pilots since 2008, 8 involving gas injection
② Borehole DAS time-lapse VSP is a real, proven 4D CO₂-monitoring tool	✅ Yes — Aquistore, Otway Stage 3, CaMI.FRS, Decatur; detection down to tens of tonnes
③ $V_P$ is “relatively sensitive” to cold CO₂ in shale	⚠️ Overstated — stiff, low-porosity rock suppresses the fluid-substitution signal

Why the intuition leaks: the rock physics

The 4D $V_P$ signal from fluid substitution is governed by the dry frame and porosity, not by the fluid alone. P-wave velocity is

$$V_P=\sqrt{\dfrac{K_{\text{sat}}+\tfrac{4}{3}\,\mu}{\rho}}$$

and in Gassmann theory saturation moves only the bulk modulus $K_{\text{sat}}$ — the shear modulus $\mu$ is unchanged. So the entire fluid-substitution $V_P$ signal rides on $\Delta K$, which caps how much velocity can move.

Crucially (Han & Batzle 2004), the fluid-saturation effect on modulus grows with porosity and approaches the Reuss bound at the critical porosity — so in stiff, low-porosity rock far from that bound, the fluid effect is small. That is the textbook reason tight rock gives a weak 4D fluid signal.

The empirical reality check

Every successful DAS / geophone 4D-VSP case sits in soft, high-porosity sandstone: Mt. Simon (~25% φ), Basal Belly River (~10% φ), Otway. Even there it is barely detectable — Decatur injected ~70,000 t of CO₂ and the time-lapse anomaly was only described as “suggestive” of CO₂ movement. If the strongest-Gassmann soft sandstone is “barely,” a stiff shale is harder a fortiori.

The “cold” nuance — right direction, modest gain (and not the only term)

Calling out cold CO₂ does grab a real physical lever, but it isn’t a free big signal:

1. Colder/denser/liquid CO₂ → higher fluid modulus $K_f$ → larger 4D signal. The Gassmann effect is roughly $\propto K_f$. Gaseous CO₂ has $K_f\!\approx\!0.02\text{–}0.05$ GPa; dense/liquid CO₂ reaches $0.1$ to $>1$ GPa (Han & Batzle 2009). But the gain is still scaled down by low porosity and is sub-proportional in $K_f$ — so modest, not transformative.
2. Don’t forget pressure. Injection-driven pore-pressure / effective-stress change often dominates the pure fluid-substitution velocity change in EOR settings.
3. Don’t forget frame weakening. CO₂ can chemically soften the rock frame; conventional Biot-Gassmann (fixed dry frame) underestimates the seismic effect, and $V_S$ drops noticeably (Creasy et al. 2024) — which argues that $V_S$ / multicomponent time-lapse may be a more sensitive 4D indicator than $V_P$ for CO₂.

Ranking of 4D effects in tight rock (qualitative): the saturation (Gassmann) term is the weakest link in stiff shale; pressure (effective stress) and the cold-CO₂ $K_f$ / thermoelastic contribution are relatively more important. Caveat: the precise pressure-vs-temperature-vs- saturation ranking in shale specifically is not closed by verified literature — it is an open question, not a settled result.

The literature gap

The direct triple — cold CO₂ + shale-oil EOR + DAS-VSP 4D $V_P$ — is essentially absent. Evidence lives in two adjacent-but-separate bodies of work:

• Tight-oil EOR field monitoring — documented Bakken pilots used no geophysical 4D method at all (rates, BHP, gas composition, logs, tracers only); CO₂ also tends to channel through high-conductivity fractures, a primary failure mode.
• DAS-VSP 4D — all success stories are in soft sandstone CO₂ storage (saline aquifers / high-φ sandstone), not shale EOR.

The closest shale-plus-fiber intersection is the Eagle Ford SOV+DAS test (Cheng et al. 2021) — but that is a preliminary feasibility result, not a cold-CO₂ 4D $V_P$ detection. Bridging the two requires rock-physics extrapolation; there is no shale field datum yet.

Annotated references

DAS / fiber 4D VSP is a proven tool

• ⭐ Egorov et al. (2017), “Time-lapse full-waveform inversion of VSP data … CO2CRC Otway,” GRL 44(14), doi:10.1002/2017GL074122 — strongest quantitative anchor: time-lapse elastic FWI resolves a small-volume CO₂ P/S velocity change from VSP.
• Pevzner, Tertyshnikov, Correa et al., “Seismic monitoring of CO2 geosequestration using multi-well 4D DAS VSP: Stage 3 of CO2CRC Otway,” IJGGC, link — 5 DAS wells track a ~15 kt plume.
• Daley/Harris/Pevzner et al. (2016), “Feasibility of time-lapse VSP monitoring at Aquistore … DAS,” IJGGC 50:248-260, link — ~27 kt visibility threshold (modeled).
• Sidenko, Tertyshnikov, Bona & Pevzner (2021), Interpretation 9(4):SJ1-SJ12, doi:10.1190/INT-2021-0038.1 — DAS sensitive enough; repeated VSP greatly reduces cost.
• Couëslan et al. (2013), ”… time-lapse 3D VSPs,” TLE 32:1268-1276, doi:10.1190/tle32101268.1 — Decatur; note it used a geophone array, not DAS; ~70 kt → only “suggestive” anomalies.

Shale + fiber (closest analog)

• ⭐ Cheng et al. (2021), “Testing of a permanent orbital surface source and DAS for monitoring of unconventional reservoirs: Preliminary results from the Eagle Ford Shale,” GEOPHYSICS 86(2):P1-P12.

Shale-oil CO₂ EOR practice

• SPE-201471 (ATCE 2020) — first Bakken/Three Forks rich-gas multiwell cyclic huff-n-puff, OnePetro.
• Energies 17(17):4200 (2024) Bakken EOR review, OSTI — 11 pilots / 8 gas; monitoring with no geophysics; fracture-channeling failure mode.

Rock physics (sensitivity / “cold” / frame weakening)

• ⭐ Han & Batzle (2004), “Gassmann equation and fluid-saturation effects on seismic velocities,” GEOPHYSICS 69(2):398-405, PDF — effect $\propto K_f$, grows with porosity, small in stiff rock.
• ⭐ Han & Batzle (2009), “CO2 Velocity Measurements and Models …,” PDF — the “cold CO₂” anchor: modulus rises as $T$ falls / $P$ rises.
• ⭐ Creasy et al. (2024), “CO2 rock physics modeling for reliable monitoring …,” Nature Comm. Earth & Env., doi:10.1038/s43247-024-01493-6 — Biot-Gassmann underestimates; $V_S$ more sensitive; prefer shear waves.

Method reference (carbonate CO₂-EOR walkaway VSP)

• SPE-222641 (ADIPEC 2024) time-lapse walkaway VSP, CO₂-WAG EOR, OnePetro.
• ADIP 2021 walkaway-VSP FWI for CO₂-EOR, OnePetro — P-velocity change detectable iff the changed cross-section exceeds FWI resolution; effective stress can dominate fluid substitution.

Open questions

• In a stiff, low-φ shale, how do saturation / pore-pressure / thermoelastic contributions to $V_P$ actually rank? Verified evidence quantifies only the saturation ($K_f$) path.
• Is there any published shale (not soft-sandstone) field/synthetic study giving a cold-CO₂ 4D $V_P$ change and a DAS-VSP detection threshold? Worth scanning recent URTeC/SPE.
• Given Creasy et al. (2024), is $V_S$ / converted-wave time-lapse the better 4D indicator for shale cold-CO₂ EOR than $V_P$?

Bottom line

DAS-VSP 4D CO₂ monitoring is mature — in soft sandstone storage. Porting it to stiff shale, the ”$V_P$ is sensitive” premise discounts, and the field literature for cold-CO₂ shale-oil EOR is a gap to be bridged by rock physics, not cited.

---

Method: this note distills a multi-source, adversarially-verified literature scan (6 search angles → 27 sources → 124 extracted claims → 25 verified by 3-vote majority, 21 confirmed). Compiled 2026-06-29.