Decoding The Black Box: The 2015 US Disaster That Revolutionized Ship Crash Investigations

The police in Mauritius investigating the grounding of the oil spill ship, the Wakashio, revealed late last week that they were unable to obtain sound from the ship’s black box. 

That is not surprising, given how ship black boxes are configured, often with multiple microphones plugged in, recording to several audio tracks simultaneously. However, this only makes the case stronger to be engaging some of the worlds’ best forensic analysts of ship black boxes, most of which resides in the US National Transport Safety Board (NTSB), an independent agency of the US Government.

The NTSB’s investigation into one of the most high-profile shipping incidents in the past decade, the sinking of the El Faro in 2015, reveal the power of technology to shine a light on ship sinkings, especially focusing on the critical evidence provided by the ship’s ‘black box’ Voyage Data Recorder.

El Faro: the deadliest shipping disaster since WW2

In 2015, an American cargo ship ended up at the bottom of the sea, taking the lives of thirty-three crew on board. It was the deadliest American maritime disaster since WW2. 

“With all our sophisticated technology, how could an accident of this magnitude happen?”

This was the question posed by Boston-based journalist, Rachel Slade, who covered the sinking and investigation into the El Faro, and turned this account into an award-winning book, Into the Raging Sea.

The El Faro wasn’t just the worst US maritime accident since WW2. The investigation and analysis of the ship’s Black Box (the Voyage Data Recorder or VDR), changed the face of modern ship investigations. It has set the benchmark for how to conduct a thorough ship incident investigation to really understand the root cause of what causes a ship to crash, spill oil or even sink.

The centerpiece of this investigation was the 510-page audio transcript from the bridge recording of the Voyage Data Recorder – the longest ever released in history.  However, when the investigators first listened to the transcript, all they could hear was howling wind.  

The full story from the NTSB is told in this 16-page illustrated summary became essential reading for any modern ship investigator and complements the 300-page Final Investigation Report that was published on 7 September 2018. 

This shows that a comprehensive incident report like that of the El Faro sinking requires 826 pages of facts, analysis, evidence and recommendations, to ensure a safe environment for global shipping – both for the environment and the 1.3 million seafarers crewing vessels around the world.

The story of how the investigators were able to forensically decipher the Black Box and extract the audio became a story in itself.

The technologies and techniques developed by the investigators at the US National Transport Safety Board (NTSB) broke new ground and the NTSB now contains some of the most advanced sound laboratories in the world to decipher ship VDRs.

What happened to the El Faro?

The El Faro was a large US container ship equipped with the latest satellite communications and advanced weather forecasting software into the world. Yet it sunk in the middle of the Bermuda Triangle, sailing straight into the eye of Hurricane Joaquin on October 1, 2015.

It later turned out that the weather data communication systems were faulty, so the captain was making routing decisions based on weather information that was 6 hours out of date.  Based on the data the captain was seeing from his weather software, he thought he was past the worst of the hurricane and was trying to sail out of the storm. In actual fact, he was heading straight into its eye.  

A subsequent lawsuit by the ships’ insurers tried to assess whether liability lay with the weather data provider, the software provider or the communications system that was being used to download accurate weather data.

The initial NTSB findings into the sinking of the El Faro had initially blamed the captain for the error of sailing into the storm. This was the usual tact of the shipping industry to ‘blame the crew’ as Jack Devanney points out in his book on shipping accidents, ‘Tankship Tromedy.’

However, the families of the El Faro insisted on finding out the truth.

This truth would never had been known had it not been an extensive mission to find, recover and then decipher data contained in the ship’s black box, the Voyage Data Recorder. 

It took three deep sea missions, and several advanced underwater rovers, before the black box could be recovered.  But that was just the beginning of the groundbreaking work, much of which relied on a variety of sophisticated computer modelling of the vessels last movements, imaging the seafloor, identifying where the VDR could be, and then extracting the relevant sounds from the VDR using complex audio data extraction software.

Through this investigation to retrieve the VDR from the bottom of the ocean, extract and decipher the data, the world of ship investigations changed forever.

It firmly established the NTSB as the world’s leading agency for conducting forensic analyses into VDRs and other root causes of ship incidents.

What is a VDR?

A VDR is like a ‘Black Box’ recorder on a plane (although it is actually colored orange). It records sounds on the bridge (the bridge is where the ship is controlled from) in a continual loop (usually a 24 hour loop), as well as takes in key data feeds from the bridge. These are then stored in a microchip within a water-tight container on the outside of a ship, to be retrieved in the event of a maritime incident or sinking.

Given the size of a ship’s bridge (a large room), there are often several microphones that are connected to the VDR to record sounds from various parts of the bridge, including ambient sound close to the ship’s communication system (VHF radio) in the event that the cable from the VHF is not directly connected into the black box recorder.

The maintenance of the VDR is critical to ship safety and is strictly regulated by the UN shipping regulator, the IMO. It is part of the ship that should be regularly assessed during regular safety assessments.

Today, no modern ship investigation would be considered complete without a full forensic analysis of the data contained within the VDR.

Unfortunately, the shipping industry investigators have lagged behind their aviation counterparts and investigations by the world’s largest domicile of ships, Panama, have been heavily criticized for being superficial, proximate (rather than root cause) and have not led to improved ship safety. Panama also only completes 50% of its reports into shipping incidents, and the risks posed by Flags of Convenience has even been raised at the UN Security Council. Given that Panama is responsible for appointing the ship assessors in many cases to certify that a ship is safe for sailing - in the case of the Wakashio, this was Class NK of Japan - serious questions would need to be asked about the thoroughness of the assessor’s role in ensuring a fully functioning VDR.

An international legal matter

The is not just an investigation matter, but a serious international legal matter.

Chapter V of the International Convention for the Safety of Life at Sea (SOLAS), regulation 20, specifies VDR carriage requirements. The system requires a minimum of the most recent 12 hours to be stored. In the event of an incident or accident, investigation authorities must be able to download and replay the VDR data without delay. Software, instructions, and special parts necessary for data extraction and replay are required to be contained within the main unit of a VDR, under IMO resolution MSC.214(81).

This is especially pertinent to vessels travelling internationally, such as the Wakashio.

Finding the VDR

The first challenge for the El Faro was to locate and extract the VDR. It took three expeditions over 11 months and the use of novel video telepresence, to be able to go down, retrieve it the VDR.

This was captivatingly explained as a whole chapter in Rachel Slade’s book Into the Raging Sea. The NTSB and Woods Hole Oceanographic Institution also highlight the lengths they went to here.

Extracting audio

Once the VDR was identified and brought up to the surface, the next phase was equally important to ensure the integrity of the audio.

Three steps were followed.

1. The NTSB invited the VDR manufacturer to witness and advise on the opening of the box. This was critical to understand the key manufacturing parts, and also for investigators to practice on dummy black boxes, before moving to the actual one.

In the case of the El Faro, the VDR was manufactured by Sperry Marine, a subsidiary of defense giant, Northrop Grumman.

2. Once the key components and the microchip was extracted, a copy was made of the audio to ensure any software did not impact or degrade the original recording. The original recording was kept safe in a separate digital data room in the event of further evidence gathering.

3. A dedicated team was convened to conduct forensic analytics on the copy of the VDR recordings, which had several sound tracks, based on how the microphones were set up.

When the NTSB team first listened to the recording, it was indecipherable. They had to take a few actions to enhance the sound. These steps are crucial to any incident investigation.

Part of this requires deep understanding of the format for how data is stored on the VDR. That is beyond the scope of this article, but details can be found here on the NMEA 0183 data string standards that uses ASCII serial communication protocols. The lack of standardization of VDR data format has been a major regulatory failing of the UN shipping agency, the IMO, despite having this issue flagged multiple times. Failure to act has cost lives and the environment, as country regulators struggle with a range of manufacturers and standards from which to extract critical ship safety data.

The six microphones of the VDR

For the El Faro, the VDR depended on input from 6 microphones. In the case of the El Faro, only two were usable. In addition, there had been significant background noise from the howling wind, the internet radio that had been turned on and the sound of a loose metal curtain ring crashing along a metal bar during the storm, next to the microphone.

The full account of what was done can be read in the first 12 pages of the full audio transcript.

There were essentially five steps

1.      Establish a dedicated forensic analytics audio team, specialized in audio extraction from VDRs and with familiarity of other potential sounds from a bridge, mechanical and software systems used on vessels. For the El Faro’s VDR, there were five audio track files that had been saved simultaneously and which audio needed to be extracted from.

2.      Invest time. It took 1000 hours over 4 months to extract the key 10 hours from the entire 26 hour recording (i.e., 100 hours of extraction effort for each usable hour of audio).  Several sections - particularly a the critical last 9 minutes - had to be listened to over 100 times, to ensure full unanimity among the eight members of the audio transcript group.

3.    Forensic analysis on the microphones. There were six microphones positioned throughout the El Faro’s bridge, according to the NTSB. Extensive digital audio filtering was necessary to enhance the audio. This was also based on simulating the effects of sound waves under such conditions, and the potential position of the microphones. NTSB and commercial enhancement software was used for this process.

Here are some of the key areas where enhancement of the audio was needed, based on the NTSB investigation.

Understanding which microphones were working

The NTSB went to significant lengths to explain which microphones were linked to which sound sources, and assessed the quality of these microphones, based on the quality of the audio tracks.

According to the NTSB report, “Extraordinary means had to be used to make some of the crew conversations intelligible. The transcript that was developed may indicate fragmented phrases and conversations and may indicate extensive passages where conversations were missing or unintelligible. This type of recording is usually caused by a combination of a high bridge noise level with a low voice signal (poor signal-to-noise ratio) or by a mechanical or electrical failure of the VDR system that severely distorts or obscures the audio information.”

Locations of the microphones matter

Based on an understanding of the location of each microphone, and how well they were working, it helped the NTSB to form a picture of what sounds were likely being picked up.

Eliminating distortion

The NTSB observed that the audio quality continually degraded with the passage of time. When compared to other investigational data, the audio quality degraded as the ship experienced deteriorating weather conditions.

Eliminating noise pollution

There were several sources of noise pollution picked up by the VDR, in particular the Satellite Radio entertainment system, as well as the clanging of a curtain against a metal rod.

Satellite radio: according to the NTSB report, “The sound of a satellite entertainment radio system (SiriusXM) was present at various times throughout the entire recording. The exact moments at which the bridge mounted satellite radio feed was in use was not specifically transcribed.” The report goes on to say that “It could not be explicitly determined if the radio was turned “on” from an off state, or its volume was raised enough to be heard on the recording. Likewise, for an “off” state, it was not explicitly clear if the radio was ever turned off or just lowered to a volume that was not detectable on the recording.”

Curtain rod: the NTSB report paid particular attention to removing noise from the metal curtain track where the microphone had been placed. In the VDR audio transcript report, the NTSB said, “The placement of the bridge mounted microphones in relation to other hardware on the ship’s bridge affected the quality of the audio recording. Specifically, microphone BMU 2A that was mounted in the vicinity of the chart table recorded repeated noise pollution that affected the group’s ability to determine the nature of conversations taking place near that microphone during critical moments of the voyage.”

4. Audio enhancement software

In order to fully decrypt the recording, specialist digital audio enhancement software were used by the transcription group, using tools built by NTSB as well as from commercial providers.

In the final transcription report, the NTSB said, “Some of the filters were NTSB in-house digital filters for speaking enhancement, while others were commercially available digital audio enhancement software programs. In some areas of the transcript, a combination of NTSB in-house digital speaking enhancement filters and commercially available filters were used simultaneously to maximize the amount of audio information available to the investigation.”

5. Audio alerts matter

It was also clear that the audio alerts heard on the bridge made a huge impact on how the crew reacted. In the final report, the NTSB said “The entire recording was filled with non-voice sounds similar to a ship being operated at sea and later in the recording, the sounds similar to a ship being operated at sea in a storm. The group determined that only sounds of bridge alarms and communications transmissions (voice or data in nature) would be added to the transcript. Any other sound that the bridge crew commented on would also be added.”

The smart use of sound in alarms in design of modern workspaces (including high pressure environments such as ship bridges where sound alerts should be better coordinated) has been highlighted by audio specialists such as LA-based entrepreneur, Frank Fitzpatrick who has talked about smarter use of sounds in hospital alerts.

Costa Concordia and the VDR

The giant cruise ship, the Costa Concordia, capsized off the coast of Italy in 2012.

Evidence obtained from forensic analytics of the VDR in the case of the Costa Concordia was critical to identify which theories to reject, and which ones were plausible for the cause of this incident.

All the critical evidence for the Costa Concordia trial was extracted from the VDR.

This allowed the investigators to triangulate orders given on the bridge with the actual movement of the rudders of the vessel, and to piece together the steps that led to the grounding and capsizing of the giant cruise ship.

It is worth noting that the Costa Concordia was built in 2005, and the Wakashio was built two years later in 2007, so it is to be expected that the Wakashio would have more modern VDR and bridge equipment available.

NTSB lessons for the Wakashio in Mauritius

The US Embassy in Mauritius had offered scientific and technical support with Wakashio oil spill on August 17.

NOAA had offered to help with scientific baselining, and Woods Hole Oceanographic Institution had offered support with the latest oil fingerprinting techniques and libraries.

Given the dedicated studios and team at NTSB, it should not be too hard to send the audio files to the US electronically for this to be analyzed as part of the investigation.

This could shed important light on the final moments on the bridge, as Mauritius continues to grapple with the worst ecological disaster in its history.