Before the Call — San Bruno Pipeline Explosion

San Bruno is a nearly perfect case study in call-description ambiguity at the opening of a major incident. Three different callers, describing the same event, told dispatch three different things: plane crash, earthquake, gas station explosion. All three were plausible. None was correct. The actual cause — a transmission pipeline rupture — was invisible to every caller because it was underground.

• Consequence description is more reliable than cause identification. What callers at San Bruno could accurately report was the consequence: a massive explosion, a fireball, structures on fire, an event felt across a wide area. That consequence description was consistent and credible across all callers, even though the cause identification varied widely. Dispatch reasoning from the consistent consequence description — large energy release event, residential area, multiple structures involved, wide impact radius — produces a more accurate operational picture than any single caller's cause identification.
• Multiple inconsistent cause reports signal scale, not confusion. When callers are reporting a plane crash and an earthquake and a gas station explosion all at the same time, from the same area, the operational signal isn't that callers are confused — it's that the event is large enough and unfamiliar enough that experienced adults with direct observation are reaching for the biggest explanatory categories they have. That's a size indicator, not a noise signal. Dispatch should weight the scale implication of that pattern.
• Infrastructure hazard recognition as a dispatch skill. Natural gas pipeline ruptures are not common, but they follow a recognizable pattern: large explosion with no visible aircraft or structural cause, fire that grows rather than diminishes with initial suppression efforts, persistent gas odor, characteristic jet-fire behavior. Dispatch centers near high-pressure transmission infrastructure benefit from pre-incident training on what a pipeline rupture looks like from a call-description standpoint — before the event, not during it.
• The incident type the dispatch center assigns at the opening shapes the entire response. If dispatch codes San Bruno as a structure fire, the response is sized for a structure fire. If dispatch codes it as an explosion with multiple structures involved and possible infrastructure origin, the response includes hazmat, utility notification, and an evacuation footprint appropriate to a gas-fed fire. The initial coding decision — made in the first minutes on incomplete information — has resource implications that propagate through the entire incident.

The 95-minute gap between PG&E's knowledge of the rupture and their first contact with 911 is the central operational failure of the San Bruno response. It is not a dispatch failure in the narrow sense — dispatch had no way to compel a call it didn't know was needed. It is a failure in the information architecture that the emergency response system depends on: the assumption that the entity with the most information about a utility emergency will share it with the people managing the public emergency.

• What utility control rooms know that dispatch doesn't. In a pipeline emergency, the utility's control room has real-time system pressure data, valve location and status, the ability to isolate sections remotely (if remote valves exist) or deploy crews to manual valves, knowledge of what's upstream and downstream of the rupture, and historical data on what that section of pipe can do. None of that information is available to dispatch from the incoming 911 calls. Dispatch operating without it is like managing a fire without knowing there's a fuel source beneath the ground keeping it alive.
• The NTSB recommendation changed the expected protocol. Following San Bruno, NTSB recommended that utility control room operators notify 911 upon suspecting a pipeline rupture. That recommendation — and subsequent regulatory frameworks — established that utility-to-911 notification is part of the emergency response system architecture, not a courtesy call. Dispatch centers in areas with transmission infrastructure should know whether their utility partners have that protocol in place, what the notification number is, and whether it has ever been tested.
• Pre-incident relationships matter more than post-incident protocol. The gap at San Bruno was partly procedural and partly cultural — PG&E employees on scene didn't know what to do, couldn't operate the valves, and the control room apparently didn't connect their system alert to a public emergency. The remedy isn't only a revised protocol document; it's a working relationship between dispatch centers and local utility emergency contacts, tested in exercises, with known names and direct numbers. The 911 call that should come in automatically in an emergency actually comes from a person who has been told to make it and knows how.
• Dispatch cannot wait for the utility call that may not come. After San Bruno, the appropriate posture for dispatch handling a large explosion in a residential area near known pipeline infrastructure is proactive — initiate contact with the utility emergency line, ask whether a system alert has been generated, request valve location and isolation status. That call may save the lives of firefighters operating in an environment fueled from below. Waiting for the utility to call first means waiting for a call that may never come.

A structure fire evacuation radius is typically defined by the fire's visible footprint plus a buffer. A pipeline rupture evacuation radius is defined by something dispatch cannot see: the pressure, the flow rate, the direction of the gas plume, and the location of the valves that will stop it. At San Bruno, the hazard zone expanded continuously for 89 minutes. The evacuation picture had to keep up with a fire that was growing, not diminishing, and whose outer boundary wasn't fixed.

• Evacuation zones must be dynamic when the fuel source is unknown. At the opening of a pipeline emergency, dispatch doesn't know the valve locations, the isolation timeline, or the rate at which the hazard zone is expanding. Initial evacuation should therefore be aggressive — establish a wide perimeter based on the visible fire plus a substantial buffer for the unseen gas plume — and plan to revise it as more information becomes available. Establishing a small initial perimeter and expanding it repeatedly is slower than establishing a large initial perimeter and contracting it if warranted.
• The evacuation and the fire suppression are in competition for the same roads. Forty-two fire agencies and 500+ firefighters converging on San Bruno needed the same access routes that evacuees needed to leave. Dispatch coordinating those two flows — inbound apparatus and outbound evacuees — on the same road network, while the hazard zone was still expanding, is a logistics problem that has to be managed in real time. Staging areas for apparatus outside the primary evacuation zone, designated inbound routes separate from outbound evacuation routes, and active communication between fire and law enforcement are the mechanisms that prevent gridlock in the response corridor.
• Re-entry decisions depend on utility confirmation, not fire appearance. At San Bruno, a fire that appears to be contained is not actually safe until the gas supply is isolated. Dispatch and IC determining re-entry eligibility based on fire visibility is dangerous — the fuel source is underground and invisible. Re-entry criteria for a pipeline rupture event should explicitly require utility confirmation that the line has been isolated and verified safe before any return to the affected area is authorized.
• Secondary explosions and reignition risk in partially evacuated areas. In the area surrounding the main rupture, residual gas in service lines, damaged structures, and the soil itself creates secondary ignition risk even after the main blaze is controlled. Dispatch managing secondary calls — structure fires, gas odor reports, reignitions — in a perimeter around the primary event should route those calls through the hazmat framework, not standard structure fire protocol, until utility clearance is confirmed for each area.

The San Bruno explosion was not a surprise in the sense that no one knew high-pressure transmission pipelines existed in residential areas of San Bruno. Line 132 ran through the neighborhood. The National Pipeline Mapping System data is publicly available. The surprise was that nobody had built a functioning emergency response relationship between the utility's control room and the dispatch center — and that gap had consequence when the gas started flowing.

• Pipeline mapping as pre-incident dispatch intelligence. The National Pipeline Mapping System (NPMS) maintained by PHMSA provides publicly accessible data on transmission pipeline locations. Dispatch centers can overlay that data with their coverage area to identify which neighborhoods, which corridors, and which intersections sit over high-pressure infrastructure. That mapping should be part of dispatch pre-incident planning, not a post-incident discovery.
• Utility emergency contact list — direct numbers, not main lines. Every dispatch center should have a direct contact list for utility emergency operations: not the PG&E customer service line, but the 24-hour pipeline emergency number that reaches a control room operator who can tell you whether a system alert has been generated and request a valve closure crew. That number should be tested annually. The contact on the other end should know what 911 needs from them in an emergency.
• The operational checklist for a large residential explosion near pipeline infrastructure. The opening moves are: (1) Dispatch fire and EMS on a structure fire + MCI framework while simultaneously (2) initiating direct contact with the utility emergency line to ask for system status, (3) broadcasting a pipeline rupture advisory to all responding units to treat the scene as a potential gas-fed fire until confirmed otherwise, (4) requesting law enforcement for evacuation perimeter management, and (5) establishing a staging area outside the potential hazard radius. None of these five steps require knowing whether it is a pipeline rupture — they are the appropriate response to a large residential explosion with unknown cause, near known transmission infrastructure.
• Post-incident protocol closure — the 5-hour reporting gap. PG&E reported the rupture to the National Response Center approximately 5 hours after the event. That delay reflects the same organizational disconnect that produced the 95-minute 911 gap. Dispatch receiving confirmation that the incident is pipeline-related should initiate the appropriate regulatory notification channels if the utility hasn't done so — NRC notification for hazardous materials releases is a federal requirement, and dispatch may be the only entity at the table tracking whether it's been done.