National Rail Incident Report

Overview

This section summarizes accepted, public-default rail incidents recorded in the national dataset.

The highest annual count in the displayed data is 24,340 in 1978. The lowest annual count is 2,376 in 2026.

Counts use accepted events marked eligible for combined public counting; Form 54 highway-rail type 07 records are excluded to avoid double counting Form 57 events.

Incident Trends

Monthly incident counts for the exact rolling 18-calendar-month window used by the explorer, from 2025-03-01 through 2026-08-31.

The highest monthly count in the displayed data is 362 in 2025-07. The lowest monthly count is 236 in 2026-04.

When Incidents Occur

Historical accepted events grouped by calendar month and day of week using the processed event date.

The highest calendar-month count in the displayed data is 41,274 in January. The lowest calendar-month count is 31,874 in April. The highest weekday count in the displayed data is 69,261 in Friday. The lowest weekday count is 44,963 in Sunday.

Calendar Month

Day Of Week

These are descriptive calendar counts only; no causal or risk conclusion is calculated.

Incident Mix Over Time

Annual counts split the combined report set into highway-rail crossing events, derailments, and other Form 54 rail-equipment events.

Across the displayed years, the report set contains 250,141 highway-rail crossing events, 129,101 derailments, and 43,606 other Form 54 rail-equipment events.

Fatalities And Injuries

Annual totals use event-level processed casualty measures available in the project database.

The highest annual fatality count in the displayed data is 1,213 in 1976. The lowest annual fatality count is 206 in 2026. The highest annual injury count in the displayed data is 5,855 in 1978. The lowest annual injury count is 511 in 2026.

When contributing FRA reports disagree on an event-level casualty value, no replacement total is fabricated; unresolved values are omitted from these sums.

Damage Distribution

Form 54 physical events grouped by their agreed documented total damage cost.

The distribution includes 172,724 Form 54 physical events with an agreed documented event-level total damage cost.

This report uses the processed Form 54 total damage cost only. It does not combine unlike Form 57 vehicle-damage measures into the same distribution.

Derailments

Accepted Form 54 events whose processed accident-type description identifies a derailment.

The highest annual derailment count in the displayed data is 8,759 in 1978. The lowest annual derailment count is 604 in 2026.

Highway-Rail Crossing Incidents

Accepted Form 57 highway-rail grade crossing events recorded in the national dataset.

The highest annual crossing-incident count in the displayed data is 13,353 in 1978. The lowest annual crossing-incident count is 1,414 in 2026.

Recent 18-Month Crossing Activity

Current open crossings grouped by the exact same rolling 18-month qualifying processed event set used by the explorer map.

Among 242,743 current open crossings in the national inventory, 239,781 have no qualifying events, 2,729 have one, and 233 have two or more in the rolling window 2025-03-01 through 2026-08-31.

The map/report invariant is 0 events = gray, 1 = orange, and 2+ = red. Counts use distinct processed physical event IDs; contributing FRA report rows are not double counted.

Repeat Crossing History

Crossings that appear in more than one accepted processed physical event across the historical report set.

47,975 crossings in the national dataset have more than one accepted, combined-count-eligible processed physical event in the historical report set.

This is a historical occurrence count, not a safety or risk rating.

Incident Causes

The most frequently recorded primary causes among Form 54 physical events for which contributing reports agree on one nonblank primary cause.

4,135 Form 54 events with multiple nonblank primary-cause values across contributing reports are not assigned an invented event-level cause.

Railroads And Operators

Physical events are counted once for each railroad represented by a contributing FRA report. One event can therefore appear under more than one railroad.

Counties And Locations

Counties with the largest numbers of accepted, combined-count-eligible events in the national dataset.

Reported Disagreements / Data Conflicts

Fields for which contributing FRA source/report records disagree on accepted individual physical events.

8,863 accepted, combined-count-eligible physical events in the national dataset have at least one stored source/report discrepancy.

A discrepancy does not mean the physical event was rejected. Conflicting source values are preserved and disclosed rather than replaced with an invented value.

Recent Rail Incidents

The most recent accepted, combined-count-eligible events in the processed dataset for the national dataset, ordered by event date.