There is a pattern in Bab that has not changed since 1985. The same injectors. The same producers. The same constant water rate, day after day, year after year. The water found the fractures early, made highways through the high-permeability streaks, and never looked back. Now the producers run 80 percent water cut, the injectors pump seawater into what might as well be a drain, and the tight matrix between the fractures—where perhaps half the remaining oil still sits—has barely been touched.
This is not waterflooding failure. It is waterflooding settling. A stable, inefficient equilibrium where viscous forces dominate, capillary forces are starved, and the thief zones have become permanent. The industry knows how to drill infill wells, convert producers to injectors, and accept rising water cut as the inevitable cost of mature field life. It is less practiced at questioning the steadiness of the flood itself, or the geometry of the pattern that was designed for a field that no longer exists.
Two complementary strategies—underutilized in Abu Dhabi but increasingly proven elsewhere—offer a different path. Unsteady water injection, pulsing and pausing injection to force water into unswept matrix, can outsmart the heterogeneity that constant flooding cannot. Smart pattern reconfiguration, evolving well patterns from inverted nine-spot to line-drive to local five-spot as fields mature, can match the flood geometry to the remaining oil. Together, they could recover significant incremental barrels from ADNOC's mature waterfloods without new drilling or major capital.
The Heterogeneity Trap
The carbonate reservoirs beneath Abu Dhabi's desert—Thamama, Arab Formation equivalents, and their deeper siblings—are among the most heterogeneous in the world. They are fractured, vuggy, layered, and streaked with high-permeability channels that act as thief zones for injected water. The matrix between these channels is tight, oil-wet, and poorly connected to the fracture network that carries most of the flow.
In this environment, conventional constant-rate water injection behaves predictably. Water channels through fractures and high-perm streaks, reaching producers in three to five years instead of the designed ten to fifteen. Early breakthrough is followed by rapid water-cut rise: 20 percent, then 40, then 60, then 80 and beyond. The incremental oil extracted per barrel of injected water declines continuously. The pattern appears to be working—pressure is maintained, voidage is replaced—but the sweep efficiency is poor, and the remaining oil is increasingly stranded.
Many of Abu Dhabi's mature onshore fields still operate well patterns designed in the 1970s, 1980s, or 1990s: inverted nine-spot arrays with peripheral injection, or early line-drive configurations that made sense when the oil column was thick and the water was fresh. Those patterns were optimized for a field with uniform permeability, no thief zones, and decades of production ahead. The field that exists today is nothing like that. Yet the pattern persists, because changing it requires capital, operational disruption, and organizational will.
The result is a quiet tragedy. Billions of barrels of original oil in place, 40 to 50 percent still in the ground, with the easiest-to-produce fraction long gone and the remaining fraction locked in matrix that the current flood will never touch. The water that sustains production is also the water that kills it, by stabilizing a bypassed, channelized flow regime.
Unsteady Water Injection—Pulsing to Improve Sweep
Unsteady water injection is the deliberate interruption of constant flow. It operates on a simple physical insight: in heterogeneous carbonates, the steady state is the enemy. When water flows continuously at constant rate, it finds the path of least resistance, stabilizes that path, and never leaves. When the rate varies—pulsing high, dropping low, shutting off entirely—the pressure transients create temporary flow reversals and capillary imbibition cycles that force water into unswept zones.
Three variants are relevant to Abu Dhabi's fields.
Cyclic injection injects at high rate for weeks, then shuts in or reduces to minimal rate for days. The pressure pulse propagates into the matrix, and the subsequent shut-in allows capillary forces to draw water into oil-wet pores that viscous flow never reached. When injection resumes, the front is broader, less channelized, and more effective at displacing oil.
Intermittent injection operates on scheduled on-off periods—two weeks on, one week off, or similar rhythms tuned to reservoir response. The longer shut-in periods allow gravity segregation and capillary redistribution to occur between fractures and matrix blocks, improving vertical sweep in layered reservoirs.
Variable-rate injection maintains continuous flow but fluctuates the rate deliberately—high for days, low for days, medium for days. This prevents the stable channel formation that constant-rate injection encourages, and mobilizes fines that can temporarily plug thief zones, diverting flow to lower-permeability regions.
The mechanism is not merely mechanical. In oil-wet carbonates, spontaneous imbibition of water into the matrix is weak under constant flow because viscous forces dominate. During shut-in periods, capillary forces have time to act. Over multiple cycles, this incremental imbibition alters local wettability, improves microscopic displacement efficiency, and gradually mobilizes oil that constant flooding left behind.
ADNOC-sponsored research has already demonstrated this in practice. A pilot test in a Middle East carbonate field showed that unsteady water injection increased oil production by roughly seven hundred barrels per day while reducing water cut by four percentage points. The mechanism was confirmed: cyclic pressure pulses improved sweep in heterogeneous zones that constant injection had bypassed for years. The limitation is scale. The pilot was a single pattern. The opportunity is field-wide implementation, enabled by the automation and real-time optimization that ADNOC's Panorama Digital Command Center and AI-driven operations can provide.
Machine learning models can now predict optimal pulse frequency, amplitude, and duration based on real-time production response. Instead of manual trial-and-error, unsteady injection becomes a continuously optimized reservoir management tool, with algorithms adjusting injector rates every hour based on producer water-cut trends, pressure transient behavior, and pattern-level balance.
Smart Pattern Reconfiguration—Evolving with the Field
Well patterns should not be static. They should evolve as fields mature, matching the flood geometry to the distribution of remaining oil. This is not a new idea in reservoir engineering textbooks. It is rarely executed in practice.
In early field life, when the oil column is thick and pressure is high, inverted nine-spot patterns or peripheral injection make sense. They quickly establish flow, maintain energy, and maximize initial production rates. This is how Bab, Rumaitha, and Asab were developed.
As water breakthrough occurs and water cut rises to 20 to 30 percent, infill drilling taps the remaining oil between existing wells. This is standard practice in Abu Dhabi and has been executed successfully in multiple fields. But infill drilling alone does not change the fundamental pattern geometry. It simply adds more wells to a configuration that is increasingly mismatched to the field's condition.
When water cut reaches 40 to 50 percent, the next logical step is converting internal producers to injectors, shifting from nine-spot to line-drive or staggered line-drive patterns. This improves areal sweep efficiency by aligning injectors and producers in rows, forcing water to sweep across the pattern rather than channeling diagonally. It also reduces the number of high-water-cut producers that must be managed, focusing production on the best remaining wells.
In late field life, when water cut exceeds 60 to 70 percent and remaining oil is isolated in pockets between channels, local five-spot or seven-spot patterns maximize sweep efficiency in specific zones. Combined with selective perforation—shutting off flooded intervals and opening remaining oil zones in middle and bottom reservoir sections—this targets the last recoverable barrels with precision rather than flooding the entire field indiscriminately.
The Abu Dhabi application is clear. Many fields have executed infill drilling but stalled at pattern conversion. The capital required for workovers, new completions, and pipeline re-routing is significant. The operational disruption of converting producers to injectors is real. And without excellent surveillance, there is risk of converting the wrong wells, creating ineffective injectors, or drilling dry holes.
But the surveillance is now excellent. 4D seismic, time-lapse saturation logs, and AI-driven history matching can map remaining oil with unprecedented precision. Pattern reconfiguration can be targeted, not blanket—only sectors with confirmed remaining oil in matrix or untapped layers receive new injectors or infill wells. This reduces capital risk and improves return on investment per dollar spent.
The Combined Approach—Unsteady Plus Smart Patterns
Unsteady injection and pattern reconfiguration are not alternatives. They are complementary tools that address heterogeneity at different scales.
Unsteady injection improves microscopic sweep efficiency—the pore-scale displacement of oil by water. It forces water into matrix blocks that constant flow bypasses, improves imbibition in oil-wet carbonates, and gradually alters local wettability through repeated cycles.
Pattern reconfiguration improves macroscopic sweep efficiency—the areal and vertical coverage of the flood. It aligns injectors and producers to match remaining oil distribution, reduces channeling by changing flow directions, and concentrates effort on high-potential sectors rather than flooding the entire field uniformly.
The recommended sequence for Abu Dhabi's mature fields is disciplined and data-driven.
First, diagnose. Use 4D seismic, production logging, and AI-driven history matching to identify sectors where water cut has plateaued or risen slowly despite constant injection, indicating bypassed oil in tight matrix or untapped layers.
Second, reconfigure. In high-potential sectors, convert selected producers to injectors, establish line-drive or local five-spot patterns, and re-perforate wells to target remaining oil zones. This is the macroscopic fix.
Third, optimize. Implement unsteady injection in the new injectors—cyclic pulses, variable rates, AI-optimized frequency based on real-time producer response. This is the microscopic fix.
Fourth, monitor. Track water cut trends, pressure transient behavior, and pattern-level balance continuously. Refine pulse parameters and pattern geometry based on data, not assumption.
Fifth, scale. Roll out to additional sectors based on pilot results, using the learning curve to reduce cost and risk with each expansion.
The economic case is compelling. Pattern reconfiguration requires two to five million dollars per sector for workovers, completions, and pipeline changes. Unsteady injection adds minimal incremental cost—automation software, variable-speed pumps, real-time monitoring infrastructure that ADNOC largely already possesses. Combined incremental recovery is conservatively five to ten percent of remaining oil in reconfigured sectors. For a field with five hundred million barrels remaining, that is twenty-five to fifty million incremental barrels at marginal cost of five to ten dollars per barrel.
The Risk—and Why It Is Worth Taking
The technical risks are specific and manageable. Unsteady injection in highly fractured zones, if poorly designed, can exacerbate rather than reduce channeling. Pulses that are too short never allow water to leave the fracture network. Pulses that are too long allow stable channels to re-establish. The mitigation is conservative initial design—two weeks on, one week off—and iterative optimization based on producer response.
Pattern reconfiguration in poorly characterized sectors risks converting productive producers to ineffective injectors, or drilling infill wells into already-swept zones. The mitigation is excellent surveillance: 4D seismic, saturation logs, and AI-driven history matching to confirm remaining oil before capital is committed. Pilot sectors with matched controls reduce risk further.
The combined risk—over-enthusiastic pulses in newly reconfigured patterns connecting to water zones or gas caps—is addressed through digital twin validation. Simulate the unsteady, reconfigured pattern in ADNOC's reservoir models before field implementation. Test the pulse design in the model, observe virtual water-cut response, refine, and only then deploy.
The economic risk is capital competition. Pattern reconfiguration and unsteady injection must compete for budget with CO₂-EOR, infill drilling, and other projects. The mitigation is phased implementation: start with unsteady injection in existing patterns, which requires minimal capital, prove the concept, then fund pattern reconfiguration in the sectors where unsteady injection alone has plateaued. This staged approach builds the business case incrementally.
A Call to Action for the Industry
Abu Dhabi's mature waterfloods are not failing. They are settling—reaching stable, inefficient equilibriums where water channels through fractures and oil remains trapped in matrix. The water that sustains production is also the water that prevents recovery of the next billion barrels. This is not a tragedy of engineering failure. It is a tragedy of engineering inertia.
Unsteady water injection and smart pattern reconfiguration offer a practical, data-driven, low-carbon path to break that inertia. No new gas infrastructure. No capture plants. No major drilling campaigns. Just intelligent manipulation of the water that is already being injected, and intelligent reconfiguration of the patterns that have outlived their original design.
ADNOC should commission three to five sector pilots in 2026, combining pattern reconfiguration with AI-optimized unsteady injection, matched against constant-injection controls in adjacent sectors. The technology is proven. The surveillance infrastructure exists. The digital optimization capability is mature. The only missing element is the organizational will to question patterns that have been static for decades.
For reservoir engineers across the Middle East and beyond, the lesson is immediate. If your waterflood has reached 60 percent water cut and plateaued, ask whether your injectors are pulsing or plodding. Ask whether your pattern was designed for the field you have today, or the field you had forty years ago. Sometimes the water that kills the well is the water that was never given a chance to sweep properly.