What a Field Is Worth Is What You Know About Where Its Fluids Are

阅读中文版 →

Strip upstream oil and gas to one sentence and it is the business of knowing where the oil is, where the gas is, and where the water went — because that is the only knowledge that turns into money. A field is worth exactly what can be said about the position of its fluids: their structure when they were trapped, and their location now, after years of being produced and pushed around. Gas matters as much as oil here, and moves more: a gas cap expands as offtake drops pressure, free gas comes out of solution once the rock falls below bubble point, and gas breaks through to a well long before anyone wanted it to. Drill where the hydrocarbon is and the well pays. Drill where it is not and millions are gone on a dry hole. Produce the zones that still hold oil and gas, leave fewer barrels behind, and recoverable reserves go up without acquiring a single new acre. Every dollar of value in this industry traces back to the quality of one picture: the subsurface, and where its fluids sit today.

Everything else is a means. The simulators — including the open-source one I help build — the seismic surveys, the history matching, the inversions: all of it is instrumentation in service of that one commercial question. None of it is the point. The point is the value of subsurface knowledge; the tools are how you earn it.

And the question itself is indifferent to flags. Hand a flow simulator a corner-point grid, a PVT table, relative-permeability curves, and twenty years of measured rates, and it solves the same mass-conservation equations whether the platform flies one flag or another. The rock cares nothing for jurisdiction, and neither does the value question — every operator on every shore is asking where the oil and gas are now, and where the water went. The same patient subsurface work meets different ground from shore to shore: on one, an aquifer-driven offshore field with no injectors anywhere, built on the commercial Petrel and Eclipse stack — a national operator’s asset, run on the same toolchain everyone uses because the toolchain follows the physics, not the border. On another, many operators and a dense service ecosystem give mature basins long second lives through monitoring, EOR, CO2, and repeated time-lapse survey. To keep the result honest, validation runs on the public Volve dataset from the North Sea, rebuilt in the open with an open-source flow simulator.

History matching is locating the value

A history match is not an academic curve-fit. It is inference about where the remaining oil and gas sit, made from almost nothing — so that the next well and the next dollar land in the right place. On an aquifer-driven field you never see the aquifer directly. You infer it from the day each well first produces water, and you read the gas the same way: the day a well’s gas-oil ratio climbs is the day the model has to explain, gas-cap drive or solution gas evolving as pressure falls. Twenty years of breakthrough dates, well by well, is the problem. Region-average properties give you starvation or flooding, never the arrival sequence the field actually recorded. And a model that reproduces the past perfectly is not a solved field — often it is a posterior that collapsed too early, a single confident story where the data only supported a cloud of them.

The commercial sting is sharper than the statistics. Production is measured only at the wells. The vast inter-well volume — including the aquifer-to-perforation pathways the whole match turns on — is inference. Between the wells, the model is blind, and between-well blindness has a price: the bypassed oil and unswept gas hiding in that volume is value left in the ground, barrels paid for in discovery and then walked past.

Seismic gives you the container, not the contents

Conventional seismic gives you structure — layers, faults, the shape of the trap. That is the container. It does not tell you the state of the fluids inside it after two decades of production. To find the value you have to see into the volume the wells cannot reach and the structural image will not resolve.

FWI: reading where the value is

This is where Full Waveform Inversion earns its keep. FWI fits the full recorded waveform — amplitude and phase — and reconstructs a high-resolution velocity and elastic field across the whole inter-well volume, beyond what ray-based methods resolve. Time-lapse, or 4D, FWI inverts the difference between a baseline and a monitor survey for the change in velocity. Production changes velocity, through fluid substitution and pressure, so that change maps where the fluid actually moved: the swept zones, the gas that came out of solution, and the bypassed oil beside them.

Read commercially, that is the sharpest instrument available for locating barrels nothing else can see. It is the one independent measurement of the swept and bypassed volume that a well-only history match can never produce — and it is exactly the map you steer an infill well by. The forward direction sets it up: feed the flow model’s saturations and pressures through a petro-elastic model — Gassmann, Batzle-Wang — and the velocity field shifts as the field produces; free gas, even a little of it, moves velocity hard, which is what makes it visible. The reservoir writes its own seismic; on Volve I close that loop into synthetic 4D. FWI runs it the other way — from real 4D back to a quantitative picture of where the value is now.

The limit is honest. FWI is expensive and ill-posed. It needs low frequencies and a good starting model. It leans on the rock-physics link to separate saturation from pressure, which the data does not hand you cleanly. And 4D FWI on the subtle changes of a quietly producing field is still frontier work. None of that makes it less the right instrument; it makes it the hard, valuable one.

Where the rock is fractured

In unconventionals, and in fractured carbonates, the natural pore plumbing is not the whole story. The fracture network is the plumbing. In stimulated rock the conductive paths are created by hydraulic fracturing, and where those fractures actually went — versus where the completion was designed to put them — decides what each stage drains and what the well produces. Fracture location is fluid-location in this rock.

The commercial weight is direct and large-dollar. Fracture geometry and position drive stage spacing and the landing zone; they set infill-well spacing; they govern parent-child interference, where a new well drains or damages an older one in a frac hit; and they decide which wells are worth a refrac. Get the fracture picture wrong and you over-drill, leave gas stranded between stages, or frac straight into a producing neighbor.

It is the same problem as before — see between and around the wells — one kind of plumbing further, and the container still comes from structural seismic. The fractures are read from microseismic during the frac, from repeated 4D and DAS fiber afterward, and increasingly from elastic and anisotropic FWI that can begin to resolve the stimulated volume and its induced anisotropy. Keep it honest: microseismic locates events, not a clean permeability map; DAS gives strain and microseismic, not a finished image; and elastic FWI for fracture characterization is emerging, frontier work, not a solved product. The rock still leaves its trace in the seismic — the reading is just harder.

What lasts

The means are fungible. A simulator open or proprietary, one seismic vendor or another, this inversion code or that — mass conservation is jurisdiction-neutral, and the tools come and go. What lasts is the picture of where the fluid is and the value pulled out of it: the reproducible result someone else can regenerate, nothing hidden in a vendor binary, knowledge that compounds into barrels over years rather than evaporating with a license.

That is also where the two shores quietly diverge. The rock is indifferent to which shore. Capital is not. The shore that turns subsurface knowledge into long-run value is the one where a twenty-year horizon reads as an asset rather than a liability — where mature basins are worth more precisely because someone keeps asking, survey after survey, frac after frac, where the leftover oil and gas are. One shore is built to keep reading that picture, year after year, and one is not.