Field Investigation and Surrogate Safety Measures

Field Investigation and Surrogate Safety Measures

Plan a purposeful field review

A field review should test diagnostic hypotheses rather than simply tour a location. Define the questions, affected users, observation conditions, measures, and team roles. If records indicate nighttime pedestrian events, a daylight inspection alone cannot resolve contrast or glare. If conflicts occur with school or bus activity, observe the relevant operating period.

Use safe observation locations and appropriate procedures. Avoid creating new hazards to collect information. Review available records and diagrams beforehand so the team knows which movements, directions, and features require attention. Include maintenance, operations, enforcement, transit, accessibility, or other expertise according to the issue.

Read physical and operational evidence

Wheel ruts, damaged delineators, scuffed barriers, tire marks, or scarred objects can suggest departures or impacts missing from reported records. They are clues, not an exact crash count. Maintenance activity, age of damage, vandalism, or other explanations may complicate interpretation. Photograph and locate the evidence, then compare it with records and staff knowledge.

Check sightlines from actual user positions, not only from a standing observer at the curb. Parked vehicles, vegetation, queues, grades, or objects can mask approaching users. A clear view at one moment may change with parking occupancy or seasonal foliage. Record the conditions and what was measured.

Observe speeds, gaps, queues, turning paths, lane choices, yielding, and crossing behavior. A vehicle-only review may miss accessibility or a route discontinuity. Walk or otherwise assess the relevant nonmotorized trip under suitable conditions, involving affected users when feasible.

Define surrogate measures

Surrogate safety measures describe interactions or conditions related to risk before enough crashes occur for a stable estimate. Examples include conflicts, evasive maneuvers, speed distributions, yielding opportunities, and queue visibility. A near-miss report can help locate an issue, but its meaning depends on the reporting or observation method.

Time to collision (TTC) estimates the time to a collision if current relative trajectories continue under the chosen model. Post-encroachment time (PET) describes the interval between one user leaving a conflict area and another arriving. They measure different aspects of an interaction and require operational definitions.

No single universal TTC below 1.5 seconds or PET below 1 second proves a future fatal crash. Thresholds depend on the method, users, geometry, and study. A low value can identify a concerning interaction, but its relationship to injury outcomes requires validation and context. FHWA Surrogate Safety Assessment Model

Control observation and measurement error

Camera view, occlusion, detection accuracy, frame rate, geometry, and classification can affect automated measures. Observers can disagree about an evasive maneuver or opportunity. Define categories, train observers, assess agreement, and review a sample of automated classifications. Do not treat software output as inherently objective and error-free.

Record denominators. Ten conflicts during ten crossing opportunities differs from ten during a thousand. A before-after reduction in conflicts may result from lower exposure rather than improved interaction. If the treatment changes route choice or demand, report that change rather than interpreting only the conflict count.

Choose comparable observation periods and conditions. Different weather, light, traffic, or camera position can change detection and behavior. Preserve original records and processing settings where authorized so another analyst can reproduce the result.

Link surrogates with crash evidence

Surrogates can support diagnosis and early monitoring, particularly for rare serious events. They can show whether intended behavior changed after a treatment. They do not automatically replace long-term crash and injury evaluation or provide a valid CMF. A conflict model calibrated for one setting may not transfer to another mode or geometry.

Use several sources when feasible. A wet-weather departure pattern, field drainage observation, surface information, and speed data can jointly support a friction-demand hypothesis. A few low-PET encounters without crash history can support further review but need interpretation of exposure and task. State the limits of the evidence.

Work a crossing investigation

Suppose residents report near misses at a crossing behind parked vehicles. Define a crossing opportunity and conflict observation method. Gather speeds, turning movements, parking conditions, approach visibility, pedestrian counts, and relevant narratives. Observe times when parking and demand are representative.

Compare conflicts by movement and condition. If some interactions occur with turning traffic while others involve approaching through vehicles, one treatment may not address both. Review daylighting, control, speed management, and accessible routes. Consult the road owner and applicable rules before selecting a change.

After a pilot, repeat observations using the same definitions and comparable conditions. Report changes in exposure, yielding, conflicts, and accessibility. A promising surrogate change supports the intended mechanism, but a short pilot does not prove the exact annual serious-injury reduction. Continue monitoring with a suitable evaluation design.

Surrogate interpretation checks

• Define the interaction, time measure, and observation method.
• Review occlusion, sample conditions, and measurement reliability.
• Connect the metric to the suspected mechanism.
• Avoid converting it into a fatal-crash effect without suitable evidence.

Summarize findings for action

A useful field report identifies observed conditions, measurements, hypotheses, evidence gaps, and possible responses. Distinguish direct observation from interpretation and avoid asserting that debris confirms a “severe recurring hazard” without context. Prioritize the most consequential findings and assign follow-up.

For the exam, select the tool that fits the hypothesis and acknowledge its limitation. Field review, narratives, diagrams, and surrogates contribute complementary information. Their value comes from a coherent investigation that supports a feasible treatment and a measurable evaluation, not merely from collecting more photographs or metrics.