Use these ranges and decisions with the approved drilling program, local well-control procedure, mud engineer input and operator policy. Lost-circulation diagnosis is a ranked field hypothesis until pressure, volume and geology evidence converge.
A high loss rate does not tell you whether the formation already contained a conductive path or the drilling process created or reopened one. Natural and induced losses can produce the same surface symptom—a falling active volume—but they imply different pressure histories, recurrence risks and prevention strategies. The defensible answer comes from converging evidence, not one convenient clue.
What “natural” and “induced” actually mean
Natural lost circulation occurs when the wellbore intersects a pre-existing conductive feature: open fractures, faults, vugs, caverns or sufficiently permeable rock. The drilling fluid gains access to a flow path that existed before the current hydraulic event.
Induced lost circulation occurs when wellbore pressure creates a new fracture, extends an existing crack, or hydraulically reopens a weak natural feature that was not conductive under the earlier pressure state. The boundary is therefore not always binary: a natural fracture can become a loss path only after ECD or a transient pressure opens it.
Do not classify by severity. Seepage can be induced, and total losses can be natural. Severity describes the observed consequence; natural versus induced describes the interpreted mechanism.
The evidence matrix
| Evidence | Natural path more likely | Induced/reopened path more likely |
|---|---|---|
| Timing | Immediate loss on entering a known fractured, vuggy or highly permeable interval. | Loss begins after a pump-rate, rheology, cuttings-loading or surge-pressure increase. |
| Pressure dependence | Substantial loss persists after circulating friction is removed. | Loss reduces or stops when authorized operating conditions lower ECD. |
| Geology | Image/log/cuttings/offset evidence supports fractures, faults, karst or depleted permeability. | A weak interval is exposed and the pressure history approaches or exceeds its integrity margin. |
| Recurrence | Offsets lose repeatedly at the same stratigraphic feature under different programs. | Recurrence tracks similar ECD, surge or operational thresholds rather than depth alone. |
| Event shape | Abrupt step on intersecting a conductive feature may occur. | A threshold response or acceleration with pressure may occur. |
No row is diagnostic alone. A natural feature can respond to pressure, while an induced fracture can continue taking fluid after pumps stop if static hydrostatic pressure remains above its closure condition.
Start with synchronized pressure and volume history
Put active volume, flow-in, flow-out, standpipe pressure, pump rate, bit/hole depth, ROP, pipe movement and PWD/ECD—when available—on one time axis. Mark transfers, additions and sensor-clock offsets. Then identify the last stable period and the first confirmed net loss.
- Did a new formation or drilling break appear immediately before the event?
- Did pump rate, mud rheology, annular loading or pipe movement change first?
- Was there a surge-prone trip or rapid pump start?
- Did the loss rate respond consistently to an authorized change in circulating condition?
- Was the weak interval newly exposed, or had it been open for hours without loss?
Why ECD matters—but does not prove causation
Equivalent circulating density combines static mud density with annular friction. Fine drilled solids, high low-shear rheology, restricted annular geometry, cuttings beds, high pump rate and some transient operations can raise the pressure seen by the formation. If this exceeds the effective integrity of the weakest exposed interval, a fracture can initiate or reopen.
But a calculated ECD is only as reliable as its inputs. It may not capture local cuttings beds, temperature/compressibility effects, tool eccentricity or transient surge. Conversely, seeing a high ECD before losses does not prove that ECD caused them; both may be consequences of another changing condition. Use PWD where available and preserve uncertainty.
Static versus dynamic behavior
If a verified loss appears only while circulating and falls materially when annular friction is removed, the evidence shifts toward a pressure-dependent mechanism. If loss continues at a comparable rate under a stable non-circulating condition, a highly conductive natural path or a fracture held open by static hydrostatic differential becomes more plausible.
This comparison must come from approved well-control and drilling procedures. It is not a recommendation to stop pumps or alter rate during an uncertain well-control event. Losses and influx can coexist, and reduced annular level can compromise hydrostatic control.
Formation evidence: use stratigraphy, not measured depth alone
Correlate the event to formation tops, lithology and structural interpretation. Carbonates may contain vugs or karst; naturally fractured intervals may show image-log or offset evidence; depleted or weak formations may fracture at lower effective pressure. Compare offsets by stratigraphic position and pressure history—not merely the same measured depth.
Surface cuttings are lagged. A lithology observed at the shakers represents an earlier downhole interval, while the pit-volume response belongs to the current hydraulic system. Align geological samples using calculated lag and align pressure/volume channels using synchronized time.
Common wrong conclusions
- “It started at the bit, so it is a natural fracture.” An older zone above the bit may have opened after ECD changed.
- “Loss stopped with pumps off, so we created a fracture.” The observation supports pressure dependence but is not unique proof.
- “The FIT value was higher than current ECD, so induced loss is impossible.” Integrity can vary with depth, depletion, stress, temperature and test interpretation.
- “Total loss means a cavern.” A sufficiently conductive fracture system can also remove all measurable returns.
- “The mud weight is unchanged, so pressure did not change.” Circulating friction and transients can change bottomhole pressure without changing static density.
Report the conclusion as a ranked hypothesis
“The event is most consistent with a pressure-dependent reopening of the weak carbonate interval below the shoe: the section circulated loss-free for 6 hours, loss began after the ECD increase, and the verified deficit reduced when the approved operating state lowered annular friction. A newly intersected natural fracture remains possible because fractured cuttings and offset losses occur in the same stratigraphic interval. Confidence: medium.”
That wording separates observation, interpretation, alternative explanation and confidence. It is more useful than declaring “induced” as a fact that the available data cannot prove.
Technical basis
Published fracture-leakage models show that loss response depends on effective aperture, fracture conductivity, pressure, time and non-Newtonian fluid behavior. Research specifically modeling drilling-induced fractures also recognizes that natural and induced fracture losses dominate many field events and require pressure-mechanical interpretation.
Common questions
Does a high loss rate prove natural fractures?
No. Both natural conductive features and induced/reopened fractures can produce severe or total losses.
Does loss only while circulating prove an induced fracture?
It strongly supports a pressure-dependent mechanism, but does not uniquely prove a new fracture. A natural weakness can open only above a circulating-pressure threshold.
What is the best field evidence?
A synchronized, repeatable relationship between net loss, operating pressure/ECD, exposed geology and depth history, with transfers and surface losses excluded.