This Grant for Rapid Response Research (RAPID) provides funding to study the San Bruno explosion and fire. On September 9, 2010, in San Bruno, California, a suburb of San Francisco, a 30-inch steel natural gas pipeline exploded in flames, igniting a fire that ultimately killed seven residents and damaged or destroyed dozens of houses. An interdisciplinary field team will provide a holistic account of the event across the engineering and social sciences by gathering data on pre-event building, topography, and weather conditions; the rate and modes of fire spread and suppression; and elements of multiorganizational coordination and decisionmaking. The aims of the project are to: (a) improve understanding of how urban fires spread and are suppressed, (b) support development and validation of next-generation urban fire simulation models, and (c) advance theories of resilience. Scholars are now at work on the next generation of fire spread modeling, employing physics-based approaches that encompass many more variables and that explicitly consider variability in building composition, contents, and urban design. As the next generation of models, however, these have lacked opportunities for validation in actual fire scenarios. Thus analysis of the San Bruno fire will appreciably advance knowledge by improving understanding of urban fire spread and suppression, and by permitting validation and refinement of existing models. Management structures such as the Incident Command System (ICS) depend on information flow amongst organizations that are familiar with its use, in situations with recognized procedures. By studying coordination where system familiarity was present, but information was sketchy and ambiguous and procedures improvised, the study will refine our knowledge of the relative strengths of command structures as tools for interorganization decisionmaking. By explicitly considering the interaction of infrastructure capacities and management systems, the project recasts critical infrastructure to include human operators and their deliberate decisionmaking and interventions.
The research team anticipates a number of broader impacts. These include the development and dissemination of improved fire-spread models for conflagrations in residential neighborhoods. These models, once validated, can inform the knowledge base of fire protection engineers and urban planners in their design of structures and spaces, as well as firefighters who may be confronted with such situations following earthquakes, landslips, or terrorist activity. Ground-level decisionmaking and coordination approaches will be bolstered by findings on the multiorganizational coordination of the event. The PIs teach numerous engineering and emergency management courses which will ensure that findings reach the user community both directly and through publications and presentations at relevant conferences and symposia. In addition, graduate students will develop skill in post-disaster fieldwork, to include engineering and social-science data gathering, analysis, and theory development. On a larger scale, recent sizable infrastructure failures, such as the collapse of the I35 bridge in Minneapolis, have highlighted the age and condition of US infrastructure. Information developed through this study may inform policymaking regarding burgeoning vulnerabilities in vital systems.
NSF RAPID: San Bruno Award Four major findings have emerged from this study. First, we have identified the new concept of "spontaneous planning," which scholars and practitioners may wish to consider in ongoing research or implement in future disasters. Spontaneous planning is distinct from normal emergency planning because it does not take place before an emergency or disaster occurs. While it is indeed an anticipatory activity, it is only witnessed after an incident occurs and takes into account actual hazards and unfolding disaster consequences. It will at times be based on or expand from existing planning documents, but it may also depart from agreed-upon procedures and result in completely unanticipated post-disaster activities. However, spontaneous planning cannot be equated with improvisation either. It is not the immediate implementation of completely unanticipated post-disaster operations, but is instead a semi-formal process of evaluating existing and unfolding problems, and determining potential solutions and required emergency management actions. It involves identifying available options and specifying additional or novel tactical guidelines based on situational awareness and the acquisition of context-specific knowledge. In short, spontaneous planning is detailed planning in the aftermath of a disaster to inform and shape improvised response and recovery activities in a more methodical (but still imperfect) manner. Second, three methods for estimation of the high radiant heat hazard area associated with a natural gas pipeline rupture were applied to the San Bruno incident and compared to each other and to observations from the incident. The models all estimate areas that are similar to the area in which houses actually burned. There is, however, uncertainty in the estimates and limitations to the analyses, especially related to representation of the effect of the initial explosion, uncertainty in wind conditions at the site, extrapolation of empirical equations, and estimation of gas release flow rate. With more validation, the methods may be useful for regional assessments of the risk associated with natural gas pipeline ruptures. Third, overall the emergency response to the San Bruno incident was quite effective, with the exception of the response by PG&E. Fire fighters from a station 300 yards northwest of the explosion arrived at the fireground within three minutes and provided an initial size up, which quickly rippled upward through mutual aid to a county-wide fire service response. The almost one hundred fire apparatus finally responding to the incident coordinated very well, rapidly and ably organizing to deal with massive radiant heat and provide exposure protection, care for victims and adaptation to damage to the local water supply network. San Francisco Fire Department foam trucks responding from San Francisco International Airport and air attack by CalFire helicopters were relatively novel tactics employed in this incident that were particularly effective in exposure protection. Police department, city leaders, San Mateo County Environmental Health Services, Calrecycle and the American Red Cross addressed traffic control, mass care, public relations, damage assessment, site security, re-entry, debris removal, environmental remediation, and long-term recovery functions. Exceptions to the successful response were the late response by PG&E to valve off the line, and less than a fully effective donations management, which exhibited characteristics of emergent improvisation rather than rational decision making during the aftermath of the incident. Fourth, there are implications that reach beyond the specifics of the incident. The gas transmission industry appears to have a large amount of aging infrastructure, the risks of which are not well understood and are not known to the general population. The industry’s physical control apparatus and emergency response capacity may not be adequate given the risks involved. Measures are needed to align the industry’s infrastructure, operating procedures and emergency response capacity consistent with the risk. A further implication of this incident is the glimpse it provides of what might occur given a major earthquake. Given one event such as this, which an entire urban county’s emergency response was able to focus on, a major earthquake such as expected in the San Francisco Bay Area or Los Angeles region could easily result in several such incidents, which would have to be largely unresponded to and which could easily lead to one or more conflagrations. In addition, three graduate students participated heavily in this project—1 PhD student in civil engineering, 1 PhD student in public administration, and 1 MPA student—and gained valuable experience in the process. The results will also be used as examples in the researchers’ courses, including Risk Analysis at the University of Delaware and Capstone in Emergency Management at the University of North Texas. Finally, the comprehensive description and analysis of the incident will be published through two journal papers, one University of Delaware Disaster Research Center report, and portions of two dissertations, all of which are currently in draft form. Results are also being shared with scholars and practitioners in academic and professional conferences.
