Engineering & Technology · Transport Infrastructure & Urban Systems

Safety evidence, maintenance, resilience and environmental impacts

A transport system is used, maintained and changed over years. Its performance includes harm, service during disruption and environmental consequences, not only the day it opens. Each outcome needs an evidence account that survives changes in use.

Review forces, structural load paths, conservation balances, graph scales and rates. The worked examples use supplied values; they are not route-design or operating specifications.

Back to the chapter

After reading, you can

Interpret safety and lifecycle evidence with explicit exposure, service and uncertainty boundaries.

A count needs an exposure and a definition

A reported crash count describes events in a defined period and area. A rate divides a compatible count by exposure, such as trips or distance travelled. Different denominators answer different questions. The count, rate and severity distribution should remain visible rather than being compressed into a single score.

A smaller count can result partly from less travel. Even a change in a comparable rate does not establish what caused it: route changes, user mix, weather, reporting and random variation may matter. A before/after comparison needs an appropriate design and uncertainty assessment before attributing the result to an intervention.

Two zero-based bars at 11 SVG units per reported crash per million trips show rates 10 before and 16 after. Supplied counts 10 and 8 are divided by 1 million and 0.5 million trips respectively. Counts fall 20%, while the observed rate rises 60%. Equal definitions are assumed, but severity, confounding, reporting and statistical uncertainty are not evaluated. A before/after ratio is not causal evidence by itself.

Build a service and harm argument

The Safe System approach described by Minnesota’s transport department recognizes human error and physical vulnerability, shared responsibility, proactive action and complementary protective elements. These principles encourage examining how the system behaves when something goes wrong. They do not certify a particular road treatment from a classroom calculation.

Maintenance changes condition over time. Evidence may include inspections, defects, observed service and the deterioration mechanisms being considered. The absence of a visible defect is not a complete demonstration of hidden condition. Connect a maintenance claim to the element, failure mode and observation quality.

Resilience concerns the service maintained or recovered under a stated disturbance. Alternative paths and backup resources can help, but a common cause may affect several at once. Distinguish a recovery target from demonstrated recovery capability. The dependency and finite-storage ideas from the preceding lessons remain relevant.

Compare environmental burdens for a comparable service

State the transport service, time horizon and system boundary before comparing environmental impacts. Construction, operation, maintenance and end-of-life effects can occur in different places and years. An apparent improvement in one stage may shift a burden to another.

A per-vehicle result cannot be converted into a per-passenger result without occupancy, and a per-tonne result needs a distance basis if journeys differ. Supplied factors can illustrate the accounting, but measured or assessed factors need their own provenance and applicability. A calculation of one emission does not establish a complete environmental preference.

Read the denominator before interpreting the trend

Construct records with identical reporting definitions: 10 reported crashes over 1 million trips before a change, and 8 over .5 million trips after it. The count falls by 2/10=20%. The observed rates are 10 and 16 reported crashes per million trips. The rate therefore rises by 6/10=60%.

This is not evidence that the change caused a worsening, nor an estimate with a demonstrated confidence interval. Severity, composition, reporting and other influences have not been examined. It shows why a claim based on the smaller count alone is incomplete.

For a separate original environmental account, transport 50 tonnes over 100 km: service is 5000 tonne-km. If a supplied boundary-specific factor is .05 kg CO₂e per tonne-km, the counted burden is 250 kg CO₂e. The factor is invented and excludes any unlisted lifecycle stages; it cannot rank real modes or projects.

Try the reasoning

Why do the count and rate move in different directions?
Show the answer and reasoning
Trips halve while reported crashes fall only 20%, so the count per trip increases.
Does a 60% observed rate increase prove a causal effect?
Show the answer and reasoning
No. The comparison lacks a causal design and uncertainty assessment.
What is missing from an unqualified “lower emissions” claim?
Show the answer and reasoning
A comparable service, applicable factors, a stated time horizon and a consistent lifecycle boundary.

What to carry forward

Keep exposure, harm, service and environmental boundaries visible throughout the lifecycle.

References

Teaching resources and research supporting this lesson’s explanations and examples.