Early in the SOFAR float program a parallel float concept started to take shape, namely the ‘one-shot’ float. While acoustically tracked floats give us tremendous detail about particle motion in the ocean, where they end up over long time will result from a combination of mean flow and eddy activity. Resolved trajectories provide a wealth of information on mesoscale activity, but their scatter makes it difficult to estimate mean flow with any accuracy. Thus, we began to explore the concept of one-shot drifters, i.e. floats that drift silently until end of mission when they make a single signal that can be used to determine their endpoint. Key to this thinking was that the float be simple and inexpensive so it could be deployed in large numbers. It is a frustrating story.
Our original idea was to design the float as an underwater clarinet. It would have a resonator tube, the length of which sets the frequency, open at one end and closed at the other with a membrane serving as the reed. The resonator tube floats thanks to another tube that provides buoyancy as well as the low-pressure cavity into which the ambient high-pressure water could squeeze past the vibrating membrane. A small explosive cap sealed the tank until it was time – at which it blew away and ambient water would rush from the resonating pipe past the membrane into the flotation. Thus, the water would flow in the opposite direction to that playing a clarinet. We built a prototype float and tested it in a high-pressure tank at ORE in East Falmouth. It worked incredibly well: when we opened the valve to let the water flow pass the membrane, it resonated like a factory whistle. Ants Leetmaa, a close friend from grad school days, coined the float the ‘abyssal whistle’. Sadly, it didn’t work when we lowered it at sea. Looking back, it became clear that it wasn’t the pipe that was resonating in the high-pressure tank, it was the tank itself!
Our next idea was to create an imploding sound by collapsing a glass sphere under pressure. Spheres are notoriously stable under pressure. To weaken a sphere so it will implode, we put a pound of thermite welding powder inside it along with a battery, a kitchen timer and magnesium wire. When the timer connects the battery to the magnesium wire it burns white-hot, hot enough to ignite the aluminum-iron oxide powder converting it into a puddle of liquid iron (melting point is 1,538°C) which would weaken the glass sphere so it implodes. We made all the preparations and lowered the glass sphere to 1000 m and listened – but all we heard was silence. When we brought it back to surface, the sphere was still intact, it hadn’t collapsed so that explained the silence. But it was completely black with soot inside! It turned out the thermite reaction did take place, producing a puddle of molten iron in the sphere. The enormous thermal contrast between the hot liquid iron and the cold water outside led to an incredible maze of cracks in the glass. But the cold water kept enough of the outer part of the glass rigid, and the high pressure kept the maze of cracks tightly compressed so there was no leakage. The reason it didn’t fall apart when we brought it back up was that by then the iron and the inner part of the glass sphere had solidified to keep it from crumbling. What a wonderful failure! For years the glass sphere was on display at WHOI and MIT.
What follows next has its roots in a 1949 note Stommel wrote in which he sketched his idea for a gondola float that would carry an array of SOFAR devices, a euphemism for a small explosive, each of which would be released from the gondola at a set time to sink. At a certain pressure it is armed (a safety precaution) and when it reaches sound channel depth it detonates. Given accurate arrival times of the explosion at a minimum of three SOFAR hydrophones one can use hyperbolic navigation to determine its position.
One day I learned from Jess Stanborough, a geophysicist at WHOI, that he had a small supply of SOFAR devices he no longer needed for his work. Thinking of Stommel’s idea I decided on plan to put each of these in a small glass sphere float with a timer to set mission length at which point the sphere would open a valve so it floods and sinks. After sinking past arming pressure, it would detonate in the sound channel. So simple, just a crystal clock count-down timer, and a mechanism for opening a valve. Don Dorson, who had worked in the GE lighting division, suggested using a pin as a valve held in place by a lever held in place by a flashbulb that would disintegrate when ignited. To keep things simple, we fabricated a small aluminum ring to sit between the glass hemispheres. The ring had a hole for the pin (sealed with an o-ring) and it supported the latch and flashbulb. The clock circuitry had only one task: to light the flashbulb at end of mission (3 weeks if I remember right). Don, I, and a senior student at Yale University (this became his capstone project) built 13 floats and ballasted them in a large tank in the basement of the Bigelow building. The director’s office was right overhead. He wasn’t too pleased to learn later that we had been handling explosives right underneath!
The floats were deployed as a cluster on a cruise out of WHOI. The SOFAR listening stations at Bermuda, Eleuthera, Grand Turk and Puerto Rico were on alert to record the arrival times from each of the 13 floats. Sadly, only one float was heard from at the expected time. That was good news, it showed that the design we had developed worked as envisioned. We can’t know what went wrong with the others, but without time or opportunity for prior testing it is hard to know what might have been the weak spots in the design.
In principle this initiative showed the viability of deploying floats for the limited purpose of determining net drift, ideally over a much longer period than just a few weeks and with a large cluster of floats to get useful statistics about mean flow and dispersion. In practice we knew already then that working with explosives was not the way forward. It would be a decade before we would return to this question.