Key takeaways
- Monitoring, an emergency stop and replacement buses enabled large evacuations, but did not give the locomotive crew enough time to escape the first strike.
- Two passenger trains were hit in the same district on 13 August, yet public evidence does not establish how Russian operators selected either target.
- Ukrzaliznytsia should review when crews may leave the cab and where trains can stop under cover. Defence planners should separately test which short-range counter-drone tools can protect moving trains.
What the evacuation counts show
On 13 August, a Russian jet-powered Shahed drone struck the locomotive of the Odesa–Dnipro passenger train in Berezivka district, killing the driver and his assistant. Ukrzaliznytsia said monitoring staff had detected the drone and ordered the train to stop near a station. The drone changed course, leaving too little time to stop and move the crew to safety. The railway reported 340 people aboard and arranged 21 buses.
Ukraine’s State Emergency Service used a different count: 317 people evacuated, including 67 children. The figures describe passengers aboard and people recorded as evacuated, so they should not be merged. The response moved hundreds of passengers away from the damaged train without further casualties. It does not show that the locomotive crew had equivalent protection.
Russia struck a second passenger train in the district later that day. Emergency services reported that 525 people, including 130 children, were evacuated; one child was injured and 17 buses deployed. The second strike confirms that passenger trains in the district faced a repeated threat that day. It does not establish what a Russian operator saw, intended or was told to target. No public forensic account has established the targeting chain, and there is no evidence that either train had a military role.
A stopped train is still a vulnerable target
The first strike exposes a limit in a procedure that otherwise worked. Stopping near a station helps move passengers to buses and organise emergency assistance. A train slowing or standing on open track remains visible, predictable and unable to manoeuvre. If a drone changes course late, the crew has less time to reach cover than passengers farther back.
A restored timetable does not record who was killed or which locomotives were damaged. In the five days before the Odesa attacks, Ukrzaliznytsia said Russia carried out 41 attacks on railway infrastructure, killing six rail workers, wounding 17 and damaging 11 locomotives. The figures cover 3–7 August, immediately before the Odesa strikes, and show why resumed service is an incomplete measure of protection. Locomotive crews, repair teams and dispatch staff absorb the human cost of keeping the network open under Russian attack.
Ukraine’s trains carry civilians between cities, support evacuations and connect the economy to European markets. Rail is also part of the EU’s Solidarity Lanes linking Ukraine with European transport corridors. Russia’s attacks therefore test both civilian mobility inside Ukraine and a transport link that Ukraine’s European partners have a direct interest in sustaining.
“The first crew had too little time to escape after the drone changed course.”
What a crew-specific review must test
The Berezivka after-action review should test whether a separate cab protocol could have increased the crew’s escape time without exposing passengers. It should establish when a crew may leave the cab, how that decision is communicated and whether any stopping points on high-risk routes offer usable cover. Ukrzaliznytsia and emergency services should publish lessons that do not reveal sensitive detection capabilities.
The strike also raises a defence question that the public evidence cannot answer: which short-range counter-drone tools can protect a moving train without disclosing fixed defensive positions or disrupting civilian operations. Ukrainian defence and railway planners should test that question against the attack timeline and local threat data before prescribing electronic warfare, guns or missile systems.
The first strike reveals a specific crew-protection gap: Ukraine’s railway moved hundreds of passengers from the two damaged trains, yet the first crew had too little time to escape between detection, stopping and impact. That judgment would change if an after-action review established a revised stop-and-shelter procedure that consistently moved crews to safety without exposing passengers, or if a field-tested defensive measure repeatedly stopped drones before they reached locomotives. Until such evidence appears, successful evacuation should be credited as a life-saving achievement, not mistaken for proof that the train itself, or the people driving it, has been protected.
Share this analysis
Mara Ionescu
Research ContributorInformation resilience, foreign influence narratives, and regional security in Romania, Ukraine, and the Republic of Moldova.
View contributor profile →
