Olive Gardens: The 2016 Andria (Italy) Train Collision

Max S
9 min readSep 19, 2021

Background

Andria is a city of 97933 people (as of January 31st, 2021) in southeast Italy, located near the shore of the Adriatic Sea, 175km/109mi east of Naples and 44km/27mi west-northwest of the coastal city of Bari.

The location of Andria in Europe.

Andria lies on the Bari-Barletta railway, a 70km/43.5mi partially double-tracked electrified branch line opening in 1965. The regular gauge (1435mm track width) railway replaced a steam-powered narrow gauge tram that previously connected the towns since 1882. In 2016 there was no uninterrupted train control or connected signaling system installed on the line, trains were dispatched individually by each station’s dispatcher after (on the single-track sections) they would get clearance from their colleague at the next station via a phone-call. In 2012 the railway had received funding from the European Union to pay for a better system, but by 2016 the project was still in the phase of collecting offers for the work on the affected section. At the time of the accident services on the line (almost exclusively regional passenger trains) were run by Ferrotramviaria (FT), a private Italian railway service provider based out of nearby Bari.

The site of the accident seen from above. Andria lies a short distance to the northwest (top left corner), Corato further away to the south (bottom right).

The trains involved

Running southbound from Barletta to Bari as ET1021 was FT ELT 200 210, a first generation Alstom Coradia Meridian made in 2008. Introduced into service in 2005 the Alstom Coradia Meridian is a three- or four-car electric multiple unit (the one involved in the accident was the four-car version) specifically developed for Italian railways. Measuring 85.5m/280ft in length the trains can reach up to 110kph/68mph despite a weight of approximately 168 metric tons* while carrying around 206 people.

*Due to a lack of information the weight is that of the very similar Coradia Continental.

FT ELT 200 210, the train providing the southbound service, photographed a month before the accident.

Travelling in the opposite direction was ET1016 from Bari to Barletta, provided on the day of the accident by FT ETR 340 341. The ETR 340 is a four-car second generation Stadler FLIRT electric multiple unit introduced in 2009. The trains weigh 120 metric tons at 74m/243ft in length and can carry 202 passengers in a two-class configuration at up to 160kph/99mph. The FLIRT-series (an abbreviation that translates to “quick light innovative regional multiple unit”) has been made since 2004, with Stadler selling over 2000 units in different configurations in almost two dozen countries.

ETR 340 341, the other train involved in the accident, photographed less than a year before the accident.

The accident

The 12th of July 2016 was a hot, humid and sunny day with temperatures reaching up to 40°C/104°F as the day approached noon. The southbound ET1021 is running almost half an hour behind schedule as it reaches Andria station at 11am, where it is scheduled to let the opposing ET1016 pass. At that time ET1016 had been waiting at Corato station just 12km/7.5mi linear distance away. Trying to sort out how to best handle the massive delay the two dispatchers presumably tried to tell each other why their local train is more important, ending with Andria’s dispatcher clearing ET1021 for departure. His colleague had apparently not quite understood the result of their conversation and released his local train too, barely a minute after ET1021 departed Andria. With ET1021 passing through Andria South station both trains were soon on a single track, heading for one another at 100kph/62mph each. The single-track line ran through endless rows of olive trees, in parts in a slight trench, meaning the drivers had next to no visibility to the left and right on straight track or far ahead in turns. The trains carried a total of 84 people, including the two train drivers and conductors.

The site of the accident seen from the perspective of ET1016 (left) and ET1021 (right).

At 11:05am ET1021 comes around a right hand curve between two olive fields, travelling at approximately 94kph/58mph, as the driver is suddenly faced with the bright yellow ET1016 right in front of him. Neither driver has any time to react, causing the trains to collide head-on at 11:05:20am at a combined speed of 195kph/121mph. The trains’ leading cars are ripped to shreds on impact, the second cars don’t fare much better. Most victims are dead within a second of the impact. ET 1016’s third car derails to the left after running into the wreckage, deflecting the remains of ET1021 off the tracks also. The rear cars of either train remain on the track, with the destruction of the forward halves of both trains having eaten up sufficient energy to stop the rear cars where they’re at. Within seconds two and a half passenger cars have turned into a chaotic pile of debris about the length of 1.5 cars, flinging pieces of debris over 30m/100ft to the side. 23 people die, 58 suffer medium or severe injuries.

Aftermath

A few minutes after the accident the first responders reach the wreckage on foot, being met by survivors crawling out of the towering pile of torn and bent steel. Soon after the first vehicles manage to reach the site of the accident a catastrophic alarm is declared, in the next few minutes over 200 responders get involved in the rescue and recovery effort. An adjacent field is used as the landing site for medical evacuation helicopters and a makeshift field hospital as ambulances can only get within a few hundred feet of the wreckage. Both involved trains were somewhat modern, but the insanely violent collision rendered any crash protection engineering or rigidity hopelessly ineffective. The leading cars of the two trains ceased to exist, going further back one of ET1016’s middle cars looks like someone cut it in half lengthwise. It is almost a miracle that most people aboard the trains survived. Originally rumors circulate that one of the drivers somehow survived, maybe jumped off the train (something that almost certainly would’ve been lethal at that speed), or that a farmer working nearby was struck and killed by a flying piece of debris. Despite the rumors being disproven they still show up in articles on the accident to this day.

ET1016’s middle cars sitting in the wreckage with severe damage.

Knowing there is no signaling system to even investigate the investigators recover one of ET1016’s data loggers, the second one and ET1021’s logger are destroyed in the collision and can’t be used. While the investigators get to work and police officers arrest both dispatchers responders keep digging through the wreckage, eventually the military is brought in to help in the recovery and identification of the victims. In some cases relatives of missing people had to come to the morgue to try and identify victims by clothing, jewelry or belongings. Both trains were modern and had no sign of a pre-accident defect, in a tragic way the standard note in the report that the brake-systems were in perfect working order is completely irrelevant this time as neither driver even got to trigger a stop before the impact. There was no technical defect, the drivers likely had no chance to do anything wrong (and weren’t even going as fast as they were allowed to).

The remains of ET1016’s middle cars, having deflected the oncoming train off the tracks.

Investigators turn their attention towards the dispatchers and the railway’s working procedures, and the longer they look into it the worse things get. First of all, Italy has a national board of rail safety (ANSF) that sets and observes the rules for the operation of a railway line in Italy. However, due to some legal loophole Ferrotramviaria wasn’t subject to the ANSF’s rules and regulations but was instead overseen by a different office who’s main purpose was the safety certification of elevators, ski lifts, tramways and cable cars. Due to this absurdity no one had noticed the ridiculous risk posed by running modern trains on a tight schedule at relatively high speeds on a single track line with no safety-system except telling dispatchers to call each other and work it out. This system had once been commonplace and sufficient, but that had been decades ago when trains made maybe half the speed and ran a few times a day, not several times an hour. Having stomached this surprise the investigators looked into the dispatchers’ records and paperwork, discovering an impressively sloppy work-ethic. It turned out that dispatchers would regularly pre-fill whole pages of the logbook, defeating the whole purpose of it except for it being filled eventually, and choosing to usually not verify which train they had let pass at sidings and stations. Completing the trifecta of ridiculous negligence was the fact that at both Andria and Corato station visitors were allowed inside the station office to ask about schedules, tickets, connecting trains or even just for random small talk. This meant that they not only had to discuss scheduling and traffic on the phone while trying to work with a useless logbook but were constantly distracted while doing so.

Prior to the accident the conversation between the two dispatchers had looked something like that (reconstruction):

  • Dispatcher at Corato: “Following train ET1642 I would send train ET1016.”
  • Dispatcher at Andria: “The track is clear for ET1016. After ET1016 arrives I would send train ET1021”
  • Dispatcher at Corato: “After train ET1016 arrives at your station the line is clear for train ET1021

ET1642 was another train on the line, scheduled to run earlier but caught up in the traffic due to the delay accumulated by ET1021. Investigators figured out that Andria’s dispatcher either forgot that he had cleared the line for ET1016 or saw ET1642 arrive, mistook it for ET1016 and allowed departure for ET1021 right into the path of ET1016. Or both. With the useless logbook and technical difficulties avoiding a meeting at Corato the dispatchers had created enough confusion for them to send the trains on a fatal collision course.

A crane picking the wreckage apart a few days after the accident.

Both trains were written off entirely after the accident, being partially broken up for scrap at the site before being hauled off on flatbeds. The damage to rolling stock alone was estimated in the official report at 9.467 million Euros/11.1 million USD. By September 2016 ANSF started the process of being placed in charge of all railways in Italy, including privately operated ones, and in October 2016 all railway lines without an automatic signaling/safety system were limited to 50kph/31mph. The FT was prohibited from reopening the affected line until they had installed an adequate safety system, which they finally got around to by 2017. Already within a month of the accident eighteen people, including the two dispatchers, are put on trial for charges surrounding negligent manslaughter and dangerous interference with rail traffic. As of September 2021 the trial against 17 of them, including the dispatchers, appears to be ongoing, with only one of the charged individuals, the head of the traffic ministry, being relieved of guilt.

Due to financial hardship worsened by the accident the Ferrotramviaria became a full subsidiary of the FS (Italian national railways) in 2017. As a side-effect of the takeover safety-procedures and oversight at the railway improved, as they now counted as part of the FS, improving the safety-level of their operations.

A firefighter climbing into the remains of ET1016.

History repeats itself

The accident had undeniable similarities to the Bad Aibling (Germany) train collision just a few months earlier, when two trains very similar to (in fact successors of) ET1016’s model collided head-on on a single-track section in Germany, claiming 12 lives. In that case an adequate safety system had been in place but been overridden by the dispatcher.

The aftermath of 2 third-generation Stadler FLIRT trains colliding at Bad Aibling in February 2016.

_______________________________________________________________

Join the discussion about this post on Reddit!

The index of all my posts on Reddit.

--

--

Max S

Train crash reports and analysis, published weekly.