At 1:30:34AM on Jun 17, 2017 the USS Fitzgerald and the container ship ACX Crystal came together just south of Yokosuka Japan. The ACX Crystal is a 730’ modern containership built in 2008 and capable of carrying 2,858 TEU of containers at a 23-knot service speed. The Fitzgerald is a $1.8B US Navy Arleigh Burke-Class Destroyer equipped with the Navy’s Aegis Ballistic Missile Defense System able to concurrently track hundreds of incoming targets up to 64 nautical miles away. Both are well-equipped ships that should be capable of avoiding others even in the very busy shipping and fishing area near Tokyo.
But, even with all the modern navigational equipment, they did come together and lives were lost. The story behind that accident is educational for anyone involved with complex systems operated by human beings and it’s frightening for those that operate small boats at sea. It’s scary how quickly well-equipped and well-crewed ships can end up coming together. For those that go to sea, it’s a reminder to be vigilant, to use all means of tracking other ships, and to be early and conservative in the decisions made to avoid other traffic. For those that build complex systems with human operators, it underlines the importance of the human interface and helping operators, especially tired operators, find the signal in all the data.
Just a bit more than 5 minutes (334 seconds) before the disaster, the ships were more than 3 miles apart. In the busy waters near Tokyo, many would consider 3 miles a comfortable distance, and in crowded waters it’s difficult to get more open water between boats. In these conditions some might even be tempted to get a coffee or check their email. It can start to feel normal. But, just 5 minutes later, 7 lives would be lost and a US Navy warship would take $367m in damage. At the time of the accident, it was the worst US naval accident in 4 decades and yet it went from near normal to disaster in only 5 minutes.
As with all disasters, there wasn’t a single mistake that lead to the loss of lives and damaged equipment. Many mistakes combined over time and these mistake accumulate and accelerate up to the accident. Clearly the Fitzgerald was operating with some unfilled positions and the ship had far too many unfixed systems problems when it got underway. But, the crew was certainly of sufficient size to successfully make the trip. Even with some systems inoperative or operating at less than full capability, this accident could have been avoided.
ProPublica recently published Fight the Ship: Death and valor on a warship doomed by its own Navy. It’s well researched and well written. They dig through 13,000 pages of naval reports to find the important facts and do an unusually good job of going through the details.
The article focuses, as it should, on the serious problems from the top of the US Navy through the Navy 7th fleet leadership to the captain of the USS Fitzgerald. Many of the same issues contributed to another collision just two months later where another 7th Fleet Destroyer also home ported at Yokosuka Japan, the USS John McCain, lost 10 lives in another collision with a merchant ship. Clearly there were serious leadership problems and these issues were covered in detail in the article. Most make sense to me and all clearly need attention.
Many of these issues were not fully within the control of the crew of the USS Fitzgerald so my focus in reading this article was more on those problems where they did have control and on the lessons that apply to all ships whether naval, commercial, or recreational. These are most relevant to me since I’m involved with both the design of complex systems that have human interfaces and also operate a small sea-going vessel.
One important lesson is a simple but important one: look out the window. In this $1.8B warship, the crew is large and responsibility is broadly distributed, and it’s easy to forget that the best way to avoid hitting something is to look out the bridge windows. Normally naval ships will have watch standers on both the port and starboard side. In this case, due to personnel shortages, the captain had made a decision to not post separate watch standers. This is a very busy waterway with large numbers of fishing vessels and considerable commercial traffic. I’ve flown over it at night on the way to Tokyo and the entire waterway is ablaze at night with lights from ships. It’s a difficult area in which to operate safely and, at the time of the accident, 3 large commercial ships and around 2 dozen smaller ships were reported to be operating in the area of the USS Fitzgerald.
It was a clear night and the nearby ships could be seen from the bridge of the Fitzgerald. It’s easy to get over-confident relying on RADARs and other navigational equipment. I’m a big believer in the utility of all this equipment, but none replace looking out the window. From experience operating a small boat at night, all shipping traffic looks closer in the dark and nothing encourages early and conservative navigational decisions more than seeing a large commercial ship only 3 miles away on a closing course. When operating a ship at 20 knots in a crowded seaway, it’s important to at least rely on visual watch-standing as one important source of anti-collision data. Nothing replaces visual watch-standing and, with a crew of over 300, this navigational safety role probably should have been prioritized higher.
The marine Automatic Identification System is perhaps my favorite navigational information source after visually scanning the horizon and using RADAR. It’s a simple system where most commercial ships broadcast their position, speed, and other operational data on an internationally assigned VHF radio channel and this data is displayed on a chart overlay showing your ship’s position and that of all ships in the area. Better navigational systems show both the current ship’s position and the closest point of approach, calculated assuming the targets maintain course and speed. This is remarkably useful data and it alone can help avoid many collisions.
The US Navy has a policy against its warships broadcasting AIS for security reasons but the naval crews still have access to the AIS data from other ships operating in their area. I argue that the USS Fitzgerald underway only 56 nautical miles from its home port would actually have increased its level of security by broadcasting AIS. Collision was, by far, their biggest security risk at that point in the mission. But it was against policy so they were not broadcasting and other ships had slightly less information to help them avoid a collision. The crew of the Fitzgerald did have this data available to them however and, after visual and RADAR, AIS is potentially the most useful and easily used data to help avoid ship-to-ship contact.
The Fitzgerald’s main mission is ballistic missile defense and one of their primary assets are large and very capable RADARs. The ship is very-well equipped. In this case, some RADARS were not working fully and the best-trained operators were not currently on board. They were operating both shorthanded and with some technical problems. Arguably their RADAR systems are designed for other types of threats quite different from “friendly” ships only 3 miles away. These complex RADAR systems in use on the Fitzgerald are not optimized for finding close-in commercial shipping and fishing traffic, but it would seem likely that even a less than fully skilled operator would have been aware of the presence of the 730’ steel containership only 3 miles away. However, I do know these RADAR systems can be difficult to use and small errors in settings can blind the crew to nearby targets.
Back in 2016, when I was out on the USS California, a nuclear powered US Navy Attack Submarine off of Florida coast (USS California Distinguished Visitors Embark), I saw an innovative step to protecting against some of these RADAR technical and operator skill shortcomings. On the California they had installed a standard, commercially available RADAR that is only used when operating on the surface and near to friendly ports.
This Virginia Class Submarine is one of the best equipped attack submarines in the world with some of the most advanced RADAR and sonar equipment available. And yet, for safe operation on the surface near ports, they augment their military systems with commercial RADAR similar to what we use on our boat Dirona. These RADARS are far easier to use than military RADARs, yet they work remarkably well and are quite reliable. Putting a commercial RADAR on the California augments their advanced weapons systems and potentially gives them more information when approaching a friendly port. Adding a good quality but inexpensive RADAR to all US Navy boats to augment their military grade systems when operating near to port would seem like a good safety decisions that could save lives and reduce equipment damage.
A US Navy Captain is considered fully responsible for the safety of their ship no matter what. Whether the captain is on the bridge or not and, even if they are not on the ship at all, if the ship hits something, they will lose their job. If I was a Navy captain, I would be tempted to install a commercial RADAR even if I had to pay for it myself (and it would be crazy for the Navy not to install it themselves).
There were many errors and poor decisions that led to this accident. Many were made at the most senior levels of the 7th fleet and the naval leaders above them, but some of the lessons apply to all ships operating at sea. Some of these on-ship lessons that stood out for me were: 1) Make sure that there is an adequate visual watch; 2) Assume the worst when nearing other vessels and take early and decisive action to avoid a collision. It’s remarkably how fast “normal” distance can become an unavoidable collision; 3) Use the RADAR. Naval RADARs do an excellent job of delivering weapons and avoiding close encounters with enemy ships, but they aren’t always excellent when operating in very close range and they are difficult to use. Adding a commercial RADAR as was done on the USS California seems like a prudent safety decision; and 4) AIS data should be used as a primary source of anti-collision data right up there with visual and RADAR. Given the risk of collision when operating near friendly ports, it probably makes sense for naval vessels to broadcast AIS data just as the commercial traffic does. This might allow a commercial crew to make better informed decisions when operating near naval vessels.
Thanks to Frank Eigler for sending me this excellent article. If you are designing complex systems with human interfaces, there are many excellent lessons to learn from this article. And, if you operate sea-going boats, this article really drives home how quickly things can go wrong and the importance of constant vigilance and early and decisive action if there is a danger of collision. It’s also a reminder that even with a big investment in navigational equipment, there is no substitute to looking out the window and being cautious.
Worth reading: Fight the Ship: Death and valor on a warship doomed by its Navy
Changes to naval procedures are written in someone’s blood. After the USS Thresher sank in 1963, the SUBSAFE Quality Assurance program was established to ensure availability of critical systems in the event of flooding. SUBSAFE governs the materials used, maintenance procedures and documentation, and the testing required to certify readiness for sea operations.
While serving as a junior officer on the USS Daniel Boone (128 crew), we visited Halifax Nova Scotia after a patrol. A pilot boarded to guide the boat into the port. He joined the Commanding Officer on the “flying bridge” next to the periscopes. It was foggy but the radar was never used because its signature said “Navy boomer” and it was under the Captain’s feet. The pilot ordered the air horn to be blown every 15 seconds. He used the echo to decide when to give the rudder order to enter the harbor. The piloting party below decks was unable to provide any assistance with bearings to local landmarks via the periscope because we couldn’t see past the bow.
Even though we are not at war must always act as if we were. That philosophy is part of the reason that the ancient radar on SSBN’s was not used. Other submarines or trawlers could be hanging off the coast intercepting signals. An SSBN could be detected and followed. Using commercial radar is a reasonable idea, just hope that the USS California’s commercial version can’t be attributed to a public Naval contract for the Virginia class. Part of a successful hunt is collecting clues to make detection easier.
When transiting to the dive point or back to port, the CO’s night orders (while asleep) stated to call him if any contact was within 4,000 yards. A lookout was with the Officer of the Deck (OOD) on the bridge. Both of them had binoculars. Radar wasn’t used then either. Container ships moved at a faster speed than the boat and inevitably a phone call to wake the Captain was needed. Junior Officers must have the courage to call even if he/she must be awakened. If the CO is tired, it is not the JO’s responsibility to worry about the sleep but rather the safety of the ship and the standing orders from the boss. Closer to land, the allowed distance of contacts was much less and a second lookout would be stationed on the fairwater planes. The CO would also be on the bridge when maneuvering in restricted waters.
One of the most scary underway times at night was near Fort Lauderdale and the Bahamas. High numbers of small craft could be at anchor, stationary, or under sail and therefore undetectable by sonar. Ascending to periscope depth carried extreme risk. It would have been helpful to have a beacon (light/sound) on the underside of these boats to aid in collision avoidance. With a high volume of traffic, such beacons might overload a submarine but seems that’s an area where improvements could be made.
We used a hand-held GPS from a nautical shop since the 30 year old submarine didn’t have one along with the piloting party taking bearings to landmark. I’ve read that Iran has been jamming GPS in the Strait of Hormuz so the question of how much technology should be deployed where is important.
After the NTSB report on the McCain collision, the Navy has announced plans to remove touch screens and revert to “physical throttles and traditional helm control”. Preventing confusion about where to find information on screens and consistency across ship classes are important for developers to consider for military user interface design. The article below also mentions cable problems with the AIS laptop system and its placement. I wonder if the submarine force will re-evaluate using joystick controllers for navigation in the wake of these lessons learned.
https://news.usni.org/2019/08/09/navy-reverting-ddgs-back-to-physical-throttles-after-fleet-rejects-touchscreen-controls
The survivability of the US Navy is questionable in a shooting war if these technology changes have made it difficult to navigate safely. How will the ships and crews perform after taking battle damage? How much technology and where should it be used? These questions are everywhere and apply to the naval fleet and automobiles as well.
I understand your reticence to use RADAR near shorelines but, at least from my perspective, the additional safety of AIS and RADAR when near shore is sufficiently high that I would like to see Navies use both when not at war even if that does given the enemy better intelligence and mean that they will now need to work extra hard to not be tracked. If you look at boomers operating in and out of the Hood Canal in Washington, they are all surfaced and they have a massive escort group to keep civilian craft away. Clearly enemies can track them on the way out of the Hood Canal and they are going to have to find a way to lose there trackers once off shore. In many ways, we already have the problem of the enemy being well informed when navy boats are operating near shore so I would use AIS and RADAR.
You have experience on board US Navy boats so know far more than I but, when looking at the technology faults and there were some, I’m concerned about them but it still seems that the primary fault was not maintaining a proper watch. The entire bridge crew can look out the window and see the container ship. In this case, a big part of the break down appears to be crew training or willingness to act.
I can’t add anything to all this great marine realm analysis. Maybe I can fill in with an aviation analogy. Here, collision avoidance is at least as important (being even more deadly), time frames shorter (considering speeds and accelerations), manpower even scarcer (small crew), one more degree of freedom in motion, and physical obstructions even worse (relatively small visible solid-angles from a normal cockpit). And yet, civil aviation authorities designate visual see-and-avoid as the primary collision-avoidance technique required of all pilots.
So, it fails regularly, at least for small airplanes, and people die.
Transponder-based collision avoidance equipment has long been standard on passenger jets, plus controllers usually keep the big boys on organized routes and away from each other. So they collide hardly ever.
AIS-analogous equipment is making its way into mandates throughout the fleet in many parts of the world. Avionics that verbally advise of collision threats are reasonably affordable to small operators, and yet, and yet … at the end of the day, these too can fail. Some bozo may operate illegally or incompetently; GPS or the electric bus may go offline. It’s still the human being, fusing together his fallible sensors with the output of the fallible automation, who has to be ready to take direct control for evasive action. Or early avoidance, better.
Hi Frank. There is no question that it’ll comes down to operators using good judgement but making it easy to do the right thing helps. I’m a huge fan of AIS and think that both marine and aviation should pick up the pace in getting adoption on transponders. Using AIS I see collision course situations often as much as 30 minutes away. That gives plenty of time to make a small adjustment and not even get close. When using eyeballs and judgement, you’ll still see the potential collision in plenty of time but it’s likely to be less far in advance and, if someone happens to be inattentive, it might only be only a few minutes away.
This was not a technology problem. They were supposed to leave in the morning. That got delayed. They didn’t get out of port until the afternoon. They still did their small boat drills and helicopter operations. All the guys on the night watch worked all day long. They had no mandatory break before their watch. They were zombies at their stations on the bridge. They had no business being on watch.
https://www.barstoolsports.com/chicago/zbt-134-what-really-happened-on-the-uss-fitzgerald
Skip to 12 minutes.
There is no question that fatigue is a big factor here and its a leading cause of most of the accidents I read about. Clearly being awake and on watch for nearly 24 hours straight is very difficult and will lead to substantial degradation of capability and many more mistakes. It’s a frequent cause of commercial fish boat accidents and perhaps the primary factor here. But there appear to also be other contributing factors in play.
With good visual, RADAR, and IS information with CPA estimates and assuming the crew felt they had the freedom to make course changes to increase safety, this accident should still have been avoidable by even a very tired crew.
I could opine at length about the skill of a typical Navy Officer of the Watch (OOW) and how normal it is for every contact with a closest point of approach (CPA) within a given limit (typically a nautical mile) to be reported to the Captain. The underlying issue here is that there can be a culture of ‘fake it till you make it’ where incompetent junior officers pretend that they know what they’re doing, and it takes a good captain to spot that and adapt accordingly.
On the more technical points…
The AN/SPS-73 is essentially a commercial nav radar (a Furuno 2120), adapted to feed video into distribution systems and track extractors (so it can be seen in the Combat Information Center [CIC] as well as on the more mundane bridge display). Things are a little different (as you saw) on a submarine, because the only time a sub uses radar is when it’s on the surface doing something like a port entry or perilous transit.
On the topic of port entries and perilous transits, the British Royal Navy has the practice of ‘Special Sea Dutymen’ (SSDs or ‘Specials’) where additional people (including numerous specialists) are placed on watch. I’d expect that the USN has similar practices, so it would have been a (bad) judgement call to determine that this transit wasn’t deserving of such attention. At specials it would be normal for the Navigating officer (rather than OOW) to have the con, and for the Captain or XO to be on the bridge.
The ‘surface’ plot in the CIC normally acts as a ‘belt and braces’ safety check against the visual picture from the bridge. So ‘surface’ would report radar contacts within a given CPA (usually wider than the CPA for reporting contacts to the Captain), and the bridge would provide visual augmentation to the CIC picture (e.g. ‘that’s a coaster’, ‘that’s a container ship’). In a complex environment like these Japanese fishing lanes the CIC would typically switch between the AN/SPS-67 long range surface radar, and the AN/SPS-73 short range navigation radar (because C band and X band are going to show you different things). The AN/SPY-1D doesn’t really have a role to play in the surface picture – it’s designed for tracking aircraft and missiles. The ‘plot’ becomes a fusion of tracks extracted from the radars (one or other or both), visual and other sources.
AIS has come along since my days at sea in the grey funnel line, but I’d expect that it too is an input to the ‘plot’, so the CIC should have had synthetic contacts that they were trying to correlate with radar track extraction. How (or even if) AIS is presented to the bridge is something outside my experience, but it wouldn’t surprise me if cheap off the shelf nav radars now have AIS overlays, whilst the systems on $Bn destroyers don’t. Unfortunately Captains can’t just buy such kit from their nearest chandler as there are all sorts of rules around ‘alterations and additions’ to control emissions signatures, radar cross section, and top weight. Also warships are terrible for having masts all over the place that create blind arcs.
The other thing that comes into play here is a mix of a culture of tech aversion, and segregation of duties. Over reliance on tech aids leads to stupid stuff happening (see ‘GPS assisted groundings’), so the bridge staff are very much told to rely on their eyeballs, compass repeaters, and doing ‘rel vel’ (relative velocity) calculations in their heads (and it’s the latter that very much separates the competent from the ‘fake it till you make it’ crowds). They may even practice the ancient arts of astronav. The CIC does all the tech stuff, as they don’t have a window to look out of (though in many cases these days they’ll have a video/IR/laser gun sight that can zoom in a lot better than any binoculars).
I also found this article from Navy Times covered things pretty well – https://www.navytimes.com/news/your-navy/2019/01/14/worse-than-you-thought-inside-the-secret-fitzgerald-probe-the-navy-doesnt-want-you-to-read/
Useful. Thanks for the additional background Chris.
The final section of the Navy Times article is particularly troubling – it indicates a stunning lack of professional specialist training for USN surface warfare officers. For context: I was a member of the (UK) Royal Naval Reserve at university many years ago; my recollection was that even in that very limited context, training on ship handling and international maritime regulations was a staple. The Royal Navy Command Exam requires a 100% pass mark on chapter and verse of the IRPCS (not just regurgitating the text, but applying specific rules to hypothetical scenarios).
A cursory bit of searching yields this 10 year old article seemingly from a USN junior officer who participated in an exchange with the Royal Navy:
https://www.mydigitalpublication.com/article/A+Rude+Awakening/100543/11210/article.html
Which seems to indicate that standardised specialist training for USN surface warfare officers was very limited at the time, with large amounts of latitude given to individual ships as to how to train their incoming JOs. Troubling.
Ah, that was a republish of a US Naval Institute article:
https://www.usni.org/magazines/proceedings/2009-01/rude-awakening
This report suggests that the UK Navy has higher standards than the US Navy on knowing COLREGS by heart. In this case it didn’t appear that COLREGS weakness was a factor in the accident and I fear I too would fail that same COLREGS test requiring 100% success to pass.
It may be the case that other forms of training like RADAR operation and general Bridge Resource Management might also not be as good as they should be. Generally, when I get a chance to speak to US Naval officers or enlisted men, I’ve been pretty impressed with the quality of their training.
Right, nor would I pass such a test. However neither of us are going to have conduct/charge of a warship crewed by several hundred people.
COLREGS familiarity was intended to be more of an example than a specific point of contention – I mentioned it because it was particularly jarring and distinct from my experience. There are interesting user experience and interpersonal relationship aspects to this incident, but they’re all in relation to things way down the list of safeguards (like CIC contacting the bridge about shipping contacts).
Assuming the 2009 article is/was accurate, it appears that the USN focuses on generalist officer training for junior warfare officers, rather than standardised training ship handling/navigation training across the fleet. I’ve not been able to find anything authoritative about centralised navigation/ship handling training for USN warfare officers since.
However, searching for “usn surface warfare stcw” leads to this:
https://www.public.navy.mil/bupers-npc/officer/Detailing/surfacewarfare/Documents/SWO%20Career%20Path%20Brief%202018.pdf
The document appears to date from 2018 – reading through it, it appears to be a proposed set of changes to the surface warfare training curriculum, including new training as well as making use of existing USCG material.
Thanks for posting your research.
You’re very welcome.
Rounding things out, because I can’t seem to leave this alone: it appears that the new career path started rolling out, with a four week OOD course being rolled out in 2019, and the full programme in place by 2021:
https://news.usni.org/2018/06/28/new-career-path-surface-warfare-officers-stresses-fundamentals-training-first-ship-time-sea
This article has some interesting context on historical changes in training for USN surface warfare officers from the 1980s to the present day.
Following up, the NTSB have published a report on the USS John S McCain collision.
The press release and accident report are here, and are worth a read:
https://www.ntsb.gov/news/press-releases/Pages/NR20190805.aspx
https://www.ntsb.gov/investigations/AccidentReports/Pages/MAR1901.aspx
Thanks for the follow-up data Murali.