From the British perspective, the ship was in a sinking condition by about 10.15 a.m. Tovey observed that ‘Bismarck was a wreck, without a gun firing, on fire fore and aft and wallowing more heavily every moment.’ Once again Müllenheim-Rechberg provides us with a vivid description from the German point of view; this time of the futility of Machinist Josef Statz trying to save a ship that was nearly fully flooded:
In the damage control centre Statz remained standing at the table. His glance fell on the damage control board, which mercilessly revealed the bad situation of the ship. Red, the colour for ‘taking water’, covered almost the entire port side; green, for ‘flooded’, showed for the port shell and powder chambers and nearly the entire starboard side – the outboard list-control tanks there had been filled for a long time.
When the journalist asked for my thoughts about the scuttling claim, I told him I had no doubt that Bismarck’s crew initiated their standard scuttling procedures when all hope was lost. Müllenheim-Rechberg indicates that this took place at about 10 a.m. Taking into account the fact that nine-minute fuses were used on the charges and that more time passed waiting for the final order to scuttle, I think the earliest time the charges placed on the cooling water intakes would have gone off was 10.15. This fits with the time the crew were seen to be abandoning ship and jumping into the sea. Kapitänleutant Gerhard Junack, who set the scuttling charges and was the last man to leave the engine room, didn’t hear them explode until he reached the battery deck just below the upper or main deck. By the time he reached the main deck, it was already awash. At 10.16, Captain Martin of the Dorsetshire described Bismarck as ‘in extremis’; in other words, at the point of death.
Bismarck was a huge, brand-new and extremely well-built ship, which was proving more difficult to sink than the British had expected. While there was no single knockout blow, as with Hood when her magazine exploded, Bismarck was definitely on her way down from the accumulation of hits she had taken from all sides. Undoubtedly some of the seawater entering her hull from 10.15 onwards would have been coming in through the fractured cooling water intakes, but it would have been a minor amount compared to the massive volumes needed to sink — and already flooding – this Goliath of a ship. It was Captain Martin who decided to put an end to the carnage by firing three Mark VII torpedoes into Bismarck’s sides: two into her starboard side at 10.20 and the final one at 10.36 into her port side. All three hit and exploded. Barely two minutes later, she heeled over quickly to port, no doubt as a reaction to the final torpedo strike, and then sank by the stern, which by this time was fully flooded.
It is impossible for me or anyone else to be absolutely sure how much the scuttling hastened Bismarck’s demise. After carefully reading all the eyewitness accounts, and having conducted a 360-degree inspection of the hull, my personal belief is that it was only by a matter of minutes and not many hours like others have speculated. A debate about who actually sank Bismarck would make for a very interesting news article if all the historical information and facts could be presented. Unfortunately, the article that appeared left no space for such valuable content and instead read like a playground argument featuring the most provocative quotes possible. Who sank the Bismarck was a fair question, and one that I tried to answer, but I did not want it to overshadow the ship’s historical significance and the effect her loss had on the naval war between Germany and Great Britain. Even though Bismarck’s operational lifespan lasted just nine days, I didn’t think that time should be defined by the final twenty-five minutes, just as I didn’t want Hood’s two-decade dominance of the high seas to be reduced to her pitiful five-minute defeat.
As we began to retrieve the Magellan, having laid a memorial plaque on the admiral’s bridge on behalf of the Kameradschaft Schlachtschiff Bismarck (Comradeship Battleship Bismarck), I was already thinking about our next, and biggest, challenge: to find the Hood. This was a search starting from scratch, with no advantages and no guarantee of success. I had delivered on my promise to Tim Gardam about locating Bismarck, but Ted Briggs and literally millions back in the UK who were following our expedition were expecting me to deliver once again. We needed to cross 1,100 nautical miles of the North Atlantic to reach the icy waters of the Denmark Strait, which gave me four and a half days to prepare the team and get ready for the hunt for Hood.
The good news from my research into the possible location for Hood was that there were plenty of positions to choose from; the bad news was that these were widely scattered, with as much as twenty nautical miles separating the furthest apart. If I simply accepted that all eight of these positions were valid, with an equal chance of being correct, it would leave me with a search box of 840 square nautical miles to cover after applying the same standard circle of error around each position. When you are out on the open ocean with nothing for your eye to use to judge distance, it is impossible to appreciate just how large an area that actually is. Even when standing on the monkey island – the highest deck of our expedition vessel – the distance to the horizon is only 6.7 nautical miles. So this was a truly huge area; four times larger than my search box for Derbyshire and nearly twice as large as that for Lucona. In terms that my sponsors could understand, I explained that it was an area greater than Paris, the largest city in Europe.
Searching for a 262-metre-long shipwreck within an area the size of urban Paris was already a daunting prospect, but it was made worse by the fact that I had lost nearly three days from my schedule due to the problems finding the hull of Bismarck. I’d started with funding from C4 for a forty-two-day expedition, including seven contingency days I could only use in case of bad weather and downtime, and arrived in the search area for Hood on the morning of day twenty-one. This meant that – excluding the seven-day contingency and the four-day transit back to Cork – I had exactly ten and a half days to find and film the wreck of Hood and to conduct the very special ceremony we had planned. My problem was that the fastest we could search the entire 840 square nautical miles, if that was necessary, was twelve days. I needed to find a way to bring the search time closer to what I had left in my schedule.
The main drawback of using a towed side-scan sonar in deep water is the excessively long time it takes to turn the ship and sonar at the end of each track-line before you can start the next. The problem is entirely due to the length of tow cable you need when searching in deep water and the correspondingly slow speed the ship has to move to minimize the drag forces on the cable. Unfortunately, speeding up is not an option because it just causes the towfish to fly higher, reducing the quality of the images, and in extreme situations can lead to the cable suddenly snapping like it did during the Derbyshire search. As I couldn’t make the search go faster, the only realistic option I had was to reduce the size of the search box by discounting one or more of the sinking positions factored into the original area.
Of the eight possible sinking positions I had to work with, there was one obvious candidate for elimination. It came from a wireless transmission sent by Admiral Liitjens to Group North,* less than an hour after sinking Hood, a message that simply read: ‘Have sunk battleship Qu. 73 AD.’ I immediately recognized Qu. 73 AD as the ‘Quadrate’ code used by the German navy during World War II to secretly report their positions at sea. I had come across such coded positions many times before in U-boat logs and had learned the procedure for converting the alphanumeric code into latitude and longitude. Basically the code represented a grid system that divided the ocean into large squares designated by an alpha reference, medium squares fifty-four nautical miles on each side designated by the first two numerals, and smaller squares six nautical miles on each side that further divided the medium squares and were designated by a second pair of numerals. Qu. 73 AD therefore stood for Quadrant AD, which was located in the Denmark Strait, and medium square 73. However, because the second pair of numerals that would normally follow the 73 to identify the smaller square was missing, the position lacked the degree of precision I needed. For some reason –
probably because his ship was still being aggressively pursued by the Prince of Wales and the cruisers Norfolk and Suffolk – Lütjens did not report the full grid reference. As this meant that Hood could have sunk anywhere within a square that measured 54 x 54 nautical miles, it was of limited use to me and I could eliminate it without any worry.
Because Lütjens’ position was in an outlying location, I could now cut the size of my search box from 840 to 591 square nautical miles. This was a very big reduction, which to my relief solved my immediate time problem in one fell swoop. I wasn’t prepared to discount any of the remaining seven positions, as I had spent several weeks in the planning phase verifying and analysing them all to make sure in my own mind that I could rely on them as wholly credible pieces of information to factor into my search box. In my experience of searching for historic shipwrecks, these seven positions from five different sources constituted a veritable bonanza of information. Generally you are fortunate to find two or three sinking positions, and the most I’d ever had in all my previous searches was four. Most importantly, because the positions came from a variety of sources, they more or less corroborated each other. While it is entirely possible for one ship to be wrong when recording its position, it is extremely unlikely that five ships navigating independently will all be similarly wrong.
All my years of research and months of analysis were now captured on a single Quarter Million Plotting Sheet† that I had reworked so many times the creases were feathered and starting to tear. Written on it in pencil and highlighter pen was all the essential navigation information I had distilled from countless documents, as well as my calculations of bearings, dead-reckoning tracks, leeway drift and circles of error. A computer is arguably a better, and prettier, place to create such a ‘re-navigation’, but I still preferred to work with pencil and paper. Old-fashioned manual plotting gives me a better feel for how the ships actually manoeuvred during the action, and it is also easier to spot errors. In the centre of the sheet was a rectangular box representing the overall search area where I was absolutely certain Hood would be found, and within it a much smaller box drawn in red. This was my high-probability box, which encircled the five sinking positions – three from Norfolk, one from Prince of Wales and one from Hood herself – that I trusted the most. If my analysis was correct, Hood’s wreck would be found somewhere in this high-probability box early during the search and all my concerns about the schedule could be forgotten.
The trick to finding the wreck quickly was to search the central high-probability box first before moving outwards to cover areas of lower probability in sequential order. This approach was based on the common-sense principle of looking in the place you think you’ve lost something first before moving on to less likely places. The concept is exceedingly simple, but when the stakes or the costs involved are very high – as in our case – sophisticated computer programs are sometimes used to determine the absolute optimum search plan with the objective of finding the target in the shortest time possible with minimal effort expended. There is even a whole field of mathematical theory on the problem that began with the research of an English statistician and Presbyterian minister called Thomas Bayes. Bayes developed the mathematical equations for his theorem in the mid eighteenth century, but they were not published until after his death and not applied in an underwater search until 1968, when John Craven of the US Navy used the methodology to locate the sunken nuclear submarine USS Scorpion.
By assigning subjective probabilities to various sinking positions or loss clues, you can use Bayesian statistics to develop a distribution map of probabilities to show you where to search first. If you fail to find the object in that location, the distribution map is updated – manually or by computer – to show you where to search next, and so on. We had used computer programs to help design the search plans for both Lucona and Derbyshire, so I had a good working knowledge of what they could do, but for Hood I opted to design my own plan. It essentially boiled down to me picking the sequence of track-lines I thought gave me the best chance of finding the wreck the quickest. As there were only ten lines in total across the entire search box, with four covering my high-probability box, it wouldn’t be hard to assign new probabilities and adjust the order of lines to run as the search progressed.
We arrived on location in the Denmark Strait at 4 a.m. on a fine, clear day with light winds and small seas: perfect weather for deep-tow operations. The temperature was noticeably colder, but the biggest change we had to get used to was the amount of daylight. The darkest it ever got was civil twilight for about four hours. I had arranged for a small transfer boat out of Iceland to meet us when we arrived, but was surprised to find a fleet of twelve fishing boats trawling the waters near the eastern edge of the search box where I wanted to begin the search. As they, like us, were going to have restricted manoeuvrability whilst towing their nets, we needed to keep a close watch on their movements to avoid the danger of them cutting across our course – or vice versa.
The boat from Iceland arrived on time and in the ship-to-ship transfer we lost the documentary director but gained a two-person Channel 4 news team, including the journalist Lindsay Taylor. Channel 4 were so pleased with the public’s response to the footage we had shot of Bismarck that they’d decided to attempt live coverage of our hunt for Hood for the nightly news, using the SeaCast satellite communication system. Having spent in excess of £100,000 to hire the SeaCast system and have it mobilized on board the Northern Horizon, they’d decided it was time to capitalize on their investment.
A few hours later, Lindsay was able to file his first story, about the deployment of the Ocean Explorer sonar officially starting the search for Hood. The fishing boats had kindly moved south, away from our location, which gave us a free run-in to the track-line I judged to have the highest probability. There was a new energy on the vessel, in part because of our new arrivals and the change in location, but mainly because of the excitement that builds with the start of every new search. Once the sonar had reached the correct depth, I had my first look at the type of terrain we’d have to contend with while trying to pick out possible targets. I’d expected the seabed to be flat, which it was, but the surface geology was more active than I had hoped. There were large areas of outcropping rock scattered about, and what appeared to be mobile sand-waves travelling across the seabed in response to a bottom current. We were mainly searching in the Irminger Basin, but this side of the search box was close enough to the lower slope of the seismically active Reykjanes Ridge for such geology to be normal. I wasn’t overly concerned by what I was seeing, but my initial thought was that it was going to be easier to look for the wreck on the western side of the box, even though the probability of that side was considerably less than the eastern side.
The reason for favouring the eastern side of the box was that four of my five most trusted sinking positions were located there in a relatively close cluster. One was based on the dead-reckoning‡ track of the Prince of Wales, fighting alongside Hood in the battle, while the other three came from Norfolk, trailing the action by fourteen nautical miles but first on the scene to report that Hood had sunk. It was the second of these wireless transmissions from Admiral Wake-Walker, commanding the First Cruiser Squadron from HMS Norfolk, that most excited me.
At 6.15 a.m., as Norfolk was steaming towards Hood’s last position, Wake-Walker sent his first transmission with what I believe was an immediate but estimated position: ‘HMS Hood has blown up in position 63° 20’ N,31° 50’ W.’ Some twenty minutes later, Hood’s remains – some balsa rafts, charred wood and hammocks bobbing about in a spreading slick of oil — could be seen from Norfolk’s bridge as she steamed past in pursuit of the enemy.
At 6.37 a.m., a second transmission was sent by Wake-Walker to the destroyers and the Admiralty: ‘HMS Hood sunk in position 63° 21’ N, 31° 47’ W. Proceed search survivors.’ This position was clearly a revision of the first, and there can be no doubt about Wake-Walker’s intentions. He was instru
cting the destroyers to steam at once to the reported position to rescue any survivors. As a cruiser that could keep up with the German ships, Norfolk could not afford to stop. Along with the Prince ’ of Wales and Suffolk, she gave the Admiralty its only chance of catching the fleeing Bismarck. This responsibility was made absolutely clear some hours later in an urgent message from the Admiralty to ‘continue shadowing Bismarck, even if you run out of fuel’. The search for survivors would be left for the destroyers, but by transmitting a second, more precise sinking position, Wake-Walker was giving anyone still alive a better chance. When I found a pink copy of this message in the archives, stamped MOST SECRET in bold red letters, I knew this was about the best pointer I could ever find to where Hood’s wreck might lie.
As the Northern Horizon progressed down the first track-line into the slightly deeper waters of the basin, the rugged geology seen in the eastern side of the box transitioned into a carpet of soft, muddy sediment. The resulting sonar images had the ideal background that shipwreck hunters like me pray for. The seabed had become a uniform tableau of burgundy against which any hard sonar targets, depicted as bright yellow, green and blue angular shapes, would readily stand out. However, other than one not terribly exciting sonar target around the halfway mark of the line, nothing else appeared. There was a moderate increase in water depth from east to west of roughly 300 metres, but that would have no real effect on the search. As long as the fine weather conditions held and I had placed my search box in the right location, I could see no impediment to finding the wreck.
In order to be awake for every moment the sonar was actively searching, my plan was to catch short sleep breaks during the turns. At the speed we were operating, the track-lines were taking nine to ten hours and the turns slightly more than five. With all the other responsibilities I had managing the expedition and doing interviews with Lindsay Taylor, I was hoping to get three hours of sleep per turn. Maintaining such a punishing schedule over the duration of a long-drawn-out search would be tough, but I was determined to be on duty when Hood popped onto my screen.
The Shipwreck Hunter Page 17