Fire Station Siting: A Euclidean Buffer Versus Network Coverage
A provincial fire department must site one new station in a city of 600,000. The standard is a 5-minute response for 90% of the urban area. The existing station covers 62% of the population within 5 minutes by road. Seven candidate sites are available: three on donated land near the ring road, two in dense older neighborhoods, and two in rapidly growing peripheral districts. The GIS team has produced a simple 3 km Euclidean buffer map that shows 88% coverage for a site near the ring road. The fire chief likes that site because the land is free. The planning department notes that the Euclidean buffer crosses a river with only two bridges and a rail yard. Traffic data show peak speeds of 18 km/h in the old core and 45 km/h on the ring road. Population is projected to grow 20% in the periphery and 2% in the core.
1. Critique the 3 km buffer analysis. Why does it overstate coverage?
Read the worked answer
The Euclidean 3 km buffer is not a response-time model. It assumes travel is possible in every direction at uniform speed, but the map crosses a river with only two bridges and a rail yard, so many points inside the circle require long detours. It also ignores the 18 km/h peak speeds in the old core versus 45 km/h on the ring road. A site near the ring road can look excellent on the map and still fail the 5-minute standard for dense neighborhoods on the other side of the river. The 88% figure is therefore an upper bound, not a coverage estimate.
The existing 62% coverage is better because it is based on road travel, but it may also be outdated. The fire chief's preference for free land is a fiscal argument, not a service argument. Donated land near the ring road may serve the periphery, where growth is projected at 20%, but it may leave the core with worse coverage.
2. Design a defensible GIS and network method for comparing the seven sites. Specify data, assumptions, and validation.
Read the worked answer
Use network analysis rather than buffers:
- Build a routable street network with turn restrictions, bridges, rail crossings, and one-way streets.
- Assign travel speeds by road class and time of day: 18 km/h in the core, 45 km/h on the ring road.
- Place the existing station and each candidate as facilities; calculate a 5-minute service area from each.
- Overlay population and projected growth at the smallest available census block or parcel level.
- Validate with at least 10 observed response times from station logs.
- Report the share of population within 5 minutes, not the share of area.
- Run sensitivity tests for peak versus off-peak and for a bridge closure.
If data are weak, report a range and state the assumptions. The output should be a coverage map plus a table of additional population covered per candidate, capital cost, and land cost.
3. How would you weigh efficiency against equity, and what siting decision rule would you recommend?
Read the worked answer
Equity matters because the core has older, denser, lower-income neighborhoods, while the periphery has growth but lower current density. A pure efficiency rule that maximizes new population covered per toman may favor the ring road. A pure need rule favors the core. Recommend a two-stage rule:
- Screen candidates by the 5-minute network standard and exclude any site that leaves more than 10% of the core outside coverage.
- Rank the remaining sites by additional population covered per cost, with a weight for projected growth.
If the ring-road donation site wins, require a smaller satellite unit in the core within five years, funded by the land saving. Publish the assumptions and validation results so the choice is auditable. Free land alone should not determine the site.