Trap geometry and prospect identification

The anatomy of a trap

A trap is a subsurface configuration that stops hydrocarbons from escaping to the surface. It needs three physical elements:

• Reservoir rock: porous and permeable, so it can both hold hydrocarbons (porosity) and deliver them to a well (permeability). Sandstones and carbonates are the two main classes.
• Seal: an overlying and laterally confining rock with effectively no permeability to hydrocarbons over geological time. Shale is the most common seal, halite the most effective; tight carbonates and anhydrite also seal.
• Geometric closure: a shape that leaves the hydrocarbons nowhere to go upward (the seal) or sideways (the reservoir is closed by its own structure, a sealing fault, or a pinchout). The deepest level that still closes is the spill point: below it, oil leaks out of the trap along the top of the reservoir and migrates on.

Traps are classified by how the closure is achieved:

• Four-way structural closure (a dome or anticline): the top of the reservoir forms a high that closes in every direction, and hydrocarbons fill it from the crest down to the spill point. No fault or pinchout is needed, so the geometry has the fewest ways to fail. Many Middle East supergiants and most drape folds over salt are four-ways.
• Three-way closure (fault-dependent): the reservoir closes on three sides and a fault closes the fourth. The column that depends on the fault exists only if the fault seals, so the fault seal becomes one more chance in the product. Fault-sealed tilted blocks and horsts are classic three-ways; Section 3.2 treats fault seal.
• Stratigraphic trap: the reservoir pinches out laterally (sand passing into shale) or is truncated at an unconformity beneath a younger seal. No structural closure is needed, but the closure is much harder to see on seismic. Turbidite fans, shoreline sands and reef buildups are the typical stratigraphic plays.
• Combination trap: structural closure with stratigraphic elements, such as a dome truncated at an unconformity, or a three-way with a sand pinchout on the fourth side. Most real traps are combinations.
• Hydrodynamic trap: a flowing aquifer tilts the oil-water contact, so oil is held on a structure too subtle to trap it in still water (Hubbert, 1953). Uncommon, but a tilted contact on a field map is its signature.

In practice, geologists name the dominant trap style, the one that carries most of the risk and most of the volume, while the actual trap is usually a blend.

Volumetrics: how much is there?

The rock that can hold oil is the reservoir inside the closure and above the contact: the gross rock volume, GRV. It is measured from the depth map of the top reservoir, the reservoir's gross thickness $H$ and the depth of the oil-water contact, and it is not the slab $A\,H$: where the column is shorter than the reservoir is thick, the base of the sand lies in water, and toward the flanks the column thins to nothing. The ratio $\mathrm{GRV}/(A\,H)$ is the geometric factor, typically between about 0.3 and 0.9. From the GRV, the Stock-Tank Oil Initially In Place is

where:

• GRV = gross rock volume above the contact, from the map (a square kilometre times a metre is $10^6$ m³, and 1 m³ is 6.29 bbl)
• $N/G$ = net-to-gross: the fraction of the gross interval that is reservoir-quality rock (0 to 1). It multiplies the gross thickness; if you start from a net pay thickness, leave $N/G$ out or you count the shale twice
• $\phi$ = effective porosity, the fraction of the net rock that is pore space (typically 0.05 to 0.35)
• $S_{\mathrm{w}}$ = water saturation, the fraction of the pore space holding water, so $1 - S_{\mathrm{w}}$ holds oil (typically 0.15 to 0.60 in oil reservoirs)
• $B_{\mathrm{o}}$ = formation volume factor: the volume a stock-tank barrel of oil occupies in the reservoir, where it is compressed and carries dissolved gas (typically 1.05 to 1.6 for black oils)

STOIIP is proportional to the GRV, $N/G$, $\phi$, $1 - S_{\mathrm{w}}$ and $1/B_{\mathrm{o}}$, but the GRV itself is not proportional to the area or the thickness: it stops growing with thickness once the base of the reservoir is below the contact everywhere, or with the contact once the trap is full to spill. Every input is uncertain: the area from the depth conversion, porosity and saturation from the logs, net-to-gross from core. A full evaluation draws each input from a distribution and reports P90, P50 and P10 volumes and their mean (Monte Carlo); the product of best guesses is only an indication, and it is not the P50 of the product.

Recoverable volume = STOIIP $\times RF$, the recovery factor: the fraction of the oil in place that can be produced, depending on the drive mechanism, well spacing and facilities. Typical oil ranges are about 5 to 30 % under solution-gas drive, 20 to 40 % with a gas cap, and 35 % or more with a strong natural water drive; waterflooding typically lifts a depletion-drive field to 30 to 50 %, and gas injection or thermal methods can take it further. Gas fields often recover 60 to 85 %, because gas expands as the pressure drops.

The five chances in detail

Exploration risk is usually split into independent elements, each with its own chance of success (Otis and Schneidermann, 1997; Rose, 2001). Count each way of failing once, under one element:

• Trap (typically 0.7 to 0.95 for a well-imaged closure): is the mapped closure real? Is the depth conversion good enough that the crest is a crest and the spill point where you put it? Low when the seismic is poor or the velocity model is suspect; high when the structure is clear in good 3D or drilled nearby.
• Reservoir (0.5 to 0.95): is there rock with enough porosity and permeability to hold and deliver oil? It depends on the depositional system, diagenesis, depth and age: well calibrated next to producing fields, poorly in frontier plays.
• Seal (0.6 to 0.95): does the top seal hold the column? Regional shales, salt and tight carbonates are the usual seals; the risk is high when the seal is thin, fractured, or breached by faults. In a fault-dependent trap the fault's own seal is a separate, sixth chance, scored once, not also under trap or top seal.
• Charge (0.3 to 0.9): is there a mature source rock that has generated enough oil or gas to fill the trap, and a migration path from the source to the trap? Charge is usually the largest risk in a frontier basin and the smallest in a mature one, where every producing field proves it.
• Timing (0.5 to 0.95): did the trap exist before the oil migrated? A trap that formed after migration is dry however good everything else is.

If the elements fail independently, their chances multiply:

with the fault seal as a sixth factor in a fault-dependent trap. Independence is an assumption, not a law: in a heavily faulted area the trap and the seal can fail for the same reason, and then the plain product misstates the risk; practitioners adjust for the correlations they know about.

Why the product is so severe. Five independent chances of 0.70 give $0.7^5 = 0.168$, a POS of 17 %; five of 0.80 give $0.8^5 = 0.328$. The POS never exceeds the weakest chance (it equals it only when every other chance is 1), and it is dominated by the weak ones: 0.9, 0.9, 0.9, 0.9 and 0.2 average 0.76, yet give 13 %. Proving one element (raising its chance to 1) multiplies the POS by the inverse of its old chance, which is why the weakest element is the one worth de-risking, and why mature basins, where most elements are already proven, give higher success rates than frontier basins.

Prospect ranking and the risk-reward trade-off

A prospect carries two numbers: the recoverable volume if it works (the success case) and its POS. Their product is the risked volume, an expected value used to rank prospects. No well ever finds the risked volume: a well is either dry or finds roughly the success-case volume. Two prospects can have the same risked volume by very different paths:

• Prospect A: 500 MMbbl if it works, POS 10 %, so 50 MMbbl risked. A giant upside with a 90 % chance of a dry hole: a "high-risk, high-reward" prospect.
• Prospect B: 100 MMbbl if it works, POS 50 %, so 50 MMbbl risked. A modest upside with an even chance of success.

The expected values are equal, but many companies prefer Prospect B:

1. Variance: independent chances of 10 % need ten wells on average for one discovery, nine of them dry; at 50 % it takes two.
2. Information value: a Prospect B discovery often opens a trend of similar prospects.
3. Portfolio impact: a program of Prospect A wells can run through its budget before its first success.

Yet Prospect A may be the only way to find a field that transforms a company, so exploration companies keep a mix: several Prospect B wells to sustain the business and a few Prospect A wells for the upside. This is a prospect inventory, and managing it is a core skill of exploration management.

Volumetric and risk pitfalls

• Using the slab instead of the map. $A\,H$ overstates the rock in a dome, badly when the reservoir is thick compared with the column. Take the area inside the contact (or the spill contour, if the contact is deeper) and integrate the rock volume from the map.
• Ignoring the spill point. A contact mapped below the spill point cannot be filled: the column stops at the spill point.
• Scoring one failure twice. Fault breach counted under both trap and seal, or migration under both charge and timing, lowers the POS for no reason. Assign each failure to one element.
• Ignoring correlation. If the reservoir and the charge depend on the same event (an organic-rich source rock and the turbidite sands deposited with it), their successes and failures are positively correlated, and the plain product is too low; it is too high only when one element's failure makes another's less likely.
• Mixing geological and commercial risk. The POS here is geological: will there be oil in the trap? Commercial risk (the oil price, fiscal terms, development cost) is tracked separately.
• Using deterministic numbers where distributions are warranted. A single STOIIP of 250 MMbbl looks precise; a P90 of 60, a P50 of 250 and a P10 of 800 MMbbl is the honest statement, and the mean of that distribution, not its P50, is what the POS should multiply.
• Overconfidence. Post-drill reviews repeatedly find that predicted chances of success exceed the success rates achieved (Rose, 2001). Calibrate your POS against your own and your basin's drilling record.
• Treating stratigraphic traps like structural ones. Their closure is harder to see on seismic, so the trap chance of a stratigraphic prospect can be 0.3 to 0.5 even when it looks good.