Integrated Reservoir Management: Decisions, Data, and Discipline

Episode 1: When Do You Act on Decision Timing Across the Reservoir Life Cycle?

Nearly 90% of upstream investment since 2019 has gone to offsetting production declines rather than supporting growth, according to the IEA's 2025 analysis of approximately 15,000 fields globally.

For an industry that has invested heavily in integrated reservoir management (IRM), the discipline designed precisely to maximise recovery and slow production decline through better decisions across the field life cycle, that is a sobering number. It suggests that despite the capability, value is still being lost somewhere between understanding the reservoir and acting on that understanding.

This is the first of a three-part series of articles that argues that the gap sits in the decision system itself. Not in the quality of the technical work, but in when decisions are made, how much the information driving them can be trusted, and whether they ever translate into field action.

When that system works, the results are measurable. Prudhoe Bay came on stream in 1977 with an initial recovery estimate of 9.6 billion bbl. By the late 1990s that had risen to 13 billion bbl, a 35% increase, not from discovering more oil but from a sequence of decisions that added recovery mechanisms the original plan had not anticipated: waterflood infill drilling, gas cap cycling, and miscible water-alternating-gas injection initiated in 1983 and expanded progressively across 70 patterns covering 21,600 acres.

Each decision had a window. Each was taken, broadly, within it (SPE 38847). The negative case is more common and rarely makes the published literature.

This series examines the decisions that shape reservoir outcomes across the full-field life cycle. The timing problem runs through all of them.

In early field life, the tension is between the cost of insufficient appraisal and the cost of running a program past the point of decision relevance. Bratvold and Begg (2010) formalised this through the Value of Information concept:

VOI = P(decision changes) × ΔValue(better decision) > Cost(data) + Cost(delay)

The time cost of delay is routinely omitted from appraisal program justifications. The drilling cost goes in. The months of delayed plateau production, discounted back to today, do not.

In mid-field life, the timing problem takes its most consequential form: pressure-maintenance initiation. The Ekofisk chalk field in the Norwegian North Sea illustrates what the reservoir cost of an implementation timeline can look like, even when that timeline carries genuine technical justification. A four-well waterflood pilot in the Tor formation ran from spring 1981 to mid-1984, with results reported as favourable by Thomas et al. (SPE 13120).

A second pilot in the Ekofisk formation was not decided until summer 1986. Full-field injection began late 1987. Voidage replacement was not achieved until 1993. The technical uncertainties were real and the facilities timeline was substantial. The published record does not support characterizing this as primarily a decision failure.

What it does demonstrate is that the reservoir did not wait.

In late 1984, seabed subsidence was discovered to have reached approximately 10 ft, prompting a major platform jacking program. Subsidence continued after injection began, driven by a water-weakening mechanism not understood until after injection started, and rates did not slow until 1998 (SPE 17852). The IRM lesson is not that the decisions were wrong. It is that the cost of the implementation timeline needs to be estimated and kept visible, whatever its causes.

In mature field life, the timing problem becomes the one most practitioners recognize. The water cut trend climbing for 6 weeks. The Hall plot slope drifting. The injection imbalance acknowledged in every review but never formally decided.

The problem is rarely detection. It is the absence of a decision boundary. Alhuthali et al. (SPE 102478) demonstrated that timely injection redistribution in developing preferential flow measurably improves sweep efficiency, with delay progressively narrowing the available corrections. The rate adjustment available early becomes a conformance treatment later.

In each phase, the structural condition is the same. The cost of action is visible. The cost of inaction is not. The full cost of delay is simply not in the room when the decision to wait is made.

Three habits close most of this gap.

State the decision explicitly before the analysis begins—not "review the injection performance." That is a study. The decision is a choice among alternatives with different consequences. Framing it this way forces the team to identify what is actually on the table.

Estimate the cost of delay, however roughly. For a pressure-maintenance decision, this means recovery efficiency as a function of initiation pressure. For a surveillance decision, it means voidage loss per each month of inaction. The habit of asking the question matters more than the precision of the answer.

Define in advance what evidence is sufficient to act. The threshold is not certainty. It is sufficiency. Ask what combination of signals, persisting for how long, with what corroboration, justifies commitment. Writing this down prevents uncertainty from becoming a permanent reason to wait.

Prudhoe Bay reached an estimated 13 billion bbl of recovery because the right decisions were taken at broadly the right times, progressively, over the producing life of the field. That result was not accidental. It was a consequence of treating each recovery opportunity as a time-sensitive decision with a cost of delay, not as a concept to be refined until certainty arrived. But acting at the right time is only half the problem.

Acting at the right time is necessary. It is not sufficient. The next episode from the IRMTS examines why.

For Further Reading

SPE 38847 Reservoir Management of the Prudhoe Bay Field by A.D. Simon and E.J. Petersen.

SPE Textbook Series Vol. 15 Making Good Decisions by R.B. Bratvold and S.H. Begg.

SPE 17852 Reservoir Aspects of Ekofisk Subsidence by R.M. Sulak and J. Danielsen.

SPE 102478 Optimal Waterflood Management Using Rate Control by A.H. Alhuthali, A. Oyerinde, and A.D. Gupta.

SPE 13120 Ekofisk Waterflood Pilot by L.K. Thomas, et al.

Muhammad Navaid Khan, SPE, is a senior specialist in reservoir engineering at ADNOC and an active leader within SPE. He chairs the SPE Integrated Reservoir Management Technical Section, co-leads the Reservoir Surveillance and Management Subcommittee under the SPE Global Reservoir Advisory Committee, and serves on the JPT Editorial Review Board. With nearly 2 decades of experience across Middle Eastern reservoirs, he leads ADNOC’s Integrated Reservoir Management Program and provides strategic direction for the field development portfolio of its offshore assets. He holds a master’s degree in petroleum engineering from Heriot-Watt University and received the 2015 SPE Middle East and North Africa Regional Service Award.

Sule Gurses, SPE, is a senior advisor in reservoir and production at SLB with 27 years of experience across operating and service companies. Her expertise spans reservoir engineering, production optimization, completions, downhole flow-control technologies, sand control, monitoring and surveillance, and integrated surface-subsurface workflows. Her work focuses on connecting subsurface, well, completion, and production perspectives to support field development, technology selection, and value-driven decisions that balance technical risk, operational feasibility, and long-term performance. Gurses serves as program chair of the SPE Integrated Reservoir Management Technical Section and as an SPE Journal reviewer, and has authored and coauthored multiple technical publications.