Crossing failures during extreme storms and flooding events can impede critical routes and severely disrupt the ability of communities to provide critical emergency services.

One scoring system was developed in the Proposed Method for Assessing the Vulnerability of Road-Stream Crossings to Climate Change: Deerfield River Watershed Pilot to identify critically important road-stream crossings based on the impacts their failure would have on emergency medical services.

This scoring focused on response times for ambulances and subsequent transport to hospitals. Network modeling was used to generate metrics. The approach essentially used scenario analysis to assess the interconnected network under current conditions and compare it to a scenario with a road-stream crossing failure. This component required collaboration with experts on emergency management and network analyses (computer scientists).

Method

First, a set of past EMS trips of ambulances and other vehicles in the study area was characterized. Each trip consisted of a starting location (i.e., address of the responding ambulance dispatch center) and a target location (e.g. the address of a patient). For each crossing location, a road closure due to culvert or bridge failure was simulated, the most time efficient alternative route was identified, and the amount of delay that each EMS trip experienced due to the failure was calculated. Then, a score was computed based on the delays.

EMS trips in the Deerfield River watershed were then synthesized using a model derived from real emergency response call data. Data were obtained from the Shelburne Communications Center for 3,144 EMS response calls to target locations in the Deerfield River watershed (Figure 5 30). These trips occurred over a five-year period, from 2011 through 2015. Based on these data, a model was created to synthesize EMS response trips throughout the entire watershed. Population density was used as a measure of the likelihood that a patient call would originate in a certain area (if more people live in an area, it is more likely that a patient call will be made there). A probabilistic distribution of patient locations throughout the watershed was created from the population density data. With a patient location, the closest ambulance center can be determined as the ambulance dispatch location and the closest hospital as the target location. Digital road maps with speed limit data were used to identify the most probable route for each EMS trip. In this way, a probabilistic distribution of EMS trips was established.

To evaluate the effect of each crossing failure, the following procedure was used:

• Calculate the shortest path lengths of all EMS trips based on the road map. Path length is based on distance and speed of travel (based on speed limits) and is a measure of time, not distance. Note the assumption that an ambulance always chooses the shortest path.
• Remove the crossing from the map (simulate failure), making that road segment impassable.
• Recalculate the shortest path lengths of all EMS trips. As a road segment becomes impassable, some shortest path lengths will increase causing a delay in EMS response. Delays of over an hour (addresses on dead end roads with no alternative routes) were truncated to 60 minutes.
• Combine delay data for all trips affected by a crossing failure to assess the impact on EMS response.
• Various metrics were computed based on the number of affected trips and the magnitude of the delays.

For more details, consult section 5.4 of the project’s report (insert link to report).