A recurring difficulty in gas explosion work is that the most important physical evidence is routinely destroyed within hours, lawfully and for good reason. Emergency crews excavate, cut out and replace the failed segment to make the area safe. By the time an investigation begins the pipe is out of the ground and the trench is backfilled. That is a real problem, but it is less often fatal to the analysis than it first appears.
Gas does not travel in straight lines
Escaping gas moves under pressure through whatever path offers least resistance, which is seldom the shortest one. Saturated soil, frozen surface layers and pavement all act as caps that force gas laterally, sometimes for considerable distances, before it finds a way upward. Seasonal conditions therefore change migration behaviour materially for the same leak in the same location.
The practical consequence is that the entry point into a building may be far from the leak, and that a building nearer the leak may be entirely unaffected while one further away fills.
Preferential pathways
Buried infrastructure creates highways. Service trenches backfilled with granular material are more permeable than surrounding native soil. Sewer and water laterals, electrical and telecommunications conduits, and abandoned service penetrations all provide low-resistance routes directly into basements and crawlspaces, frequently through the annular space around the pipe rather than the pipe itself.
Mapping those pathways is a substantial part of the work. Utility as-builts, one-call locate records and permit history establish what is buried where, and the entry point into the structure usually identifies which pathway to examine first.
Establishing the leak source without the pipe
The removed segment is usually retained by the operator, and its condition is highly informative. Metallurgical examination distinguishes among the dominant leak causes: external corrosion with characteristic pitting and coating disbondment, stress corrosion cracking with its distinctive colony morphology, mechanical damage from third-party excavation, and joint or fitting failure.
Fracture surface examination can indicate whether a failure was progressive or sudden, and whether damage preceded the incident. Where the segment has been cut for removal, the cut faces are obvious and separable from the failure itself, provided the examination is done before the sample is further handled.
Corrosion and third-party damage tell different stories
The two most common causes point at different parties and different records. External corrosion implicates coating condition, cathodic protection adequacy and the operator's inspection and survey programme, all of which generate documentation. Third-party damage implicates locate accuracy, excavation practice and whether a one-call ticket was placed and honoured.
Damage is also frequently not immediate. A pipe struck during excavation may leak weeks or months later, which makes locate and excavation records from well before the incident relevant.
Migration studies after the fact
Where the question is whether a given path could have carried gas to a given building, that can often be tested directly. Bar-hole surveys measure gas concentration in soil at depth across a grid. Tracer gas introduced at the repaired location and monitored along candidate paths demonstrates connectivity. Soil permeability testing supports or undermines a claimed migration distance.
These are ordinary field techniques, but they are considerably more persuasive than argument from first principles, and they remain available after the pipe is gone.
The records that build the timeline
Leak survey history for the segment, prior odour complaints, cathodic protection readings, pressure telemetry and work orders together establish how long a condition may have existed and what the operator knew. Where a leak was previously graded and scheduled rather than repaired, the grading criteria and the operator's own procedures become central.
Retention periods vary and are often short. A preservation request should cover the removed pipe segment itself, which is otherwise liable to be scrapped as ordinary material.
Season changes the answer
Migration behaviour is strongly seasonal, which is why an incident cannot be assessed against conditions at the time of inspection. A frozen surface layer or saturated ground caps the soil and forces gas laterally and further; dry, permeable, unfrozen ground lets it vent harmlessly upward over a short distance. The same leak that dissipated without incident for months can fill a basement after a hard freeze or a heavy rain.
Reconstructing the conditions therefore means reconstructing the weather: frost depth, precipitation, groundwater elevation and ground temperature for the relevant period, all of which are obtainable from public records. Where a leak is known to have pre-existed the incident, this analysis often explains why it became dangerous when it did.
Sequencing the analysis
The reconstruction that holds together generally runs in one direction: establish the entry point into the structure from the damage and the building examination, identify the candidate pathways serving that entry point, test connectivity along them, and only then work back to the source and its cause. Beginning from an assumed source and reasoning forward tends to find what it set out to find.
That ordering also makes the analysis auditable. Each step rests on evidence that can be shown independently of the conclusion, which is the property that distinguishes a reconstruction from a narrative.
This article is general technical orientation, not a failure analysis, an engineering opinion, or advice on any specific matter. Determining the cause of a particular incident requires hands-on examination by a credentialed expert.