This section is a summary of the technical challenges of seabed seismic experiments during the early years of the science, broadly between the first seismic experiments at sea in 1935, and the end of the Bullard laboratories involvement in Marine Seismic technology in 2000.
At the start of seismic measurements at sea it was assumed that in order to get results you needed to emulate land seismic experiments, replacing the land surface with the seabed, so that you needed to put your geophone sensor directly on the seabed, and similarly fire you shots on the seabed. The early land geophones were simply single component vertical motion sensors and it was assumed that the seabed sensor should be similar to land geophones, only waterproof.
Tom Gaskell deploying a pressure proof land geophone in 1939
Diagram of Geophone used on HMS Jason in 1938 – a land geophone in a pot!
When these were deployed from the ship it became clear that it was difficult to ensure that they landed upright, so an early iteration produced a flat geophone that would operate whichever way up it landed. This presents a technical challenge as the geophones used a mass on a spring suspension, with the static spring tension supporting the mass against the force of gravity, and when the geophone was inverted this force was reversed, so that the geophone needed to be designed to work with a wide range of mass positions. This can only be achieved in a simple mechanical system by using strong spring suspension so that whatever position the geophone lands, the mass will still be suspended and won’t reach the limits of its travel. A direct consequence of using a stronger spring for a given mass is that the resonant frequency of the spring-mass combination is higher. Since spring-mass geophones are only sensitive to frequencies above the resonant frequency, a simple geophone designed to be used in any orientation does not record frequencies below about 5 to 10 Hz.
This was an early attempt to produce a reversible geophone – Much later work showed that this was very near the optimum physical shape for a geophone, but subsequent designs moved away from this configuration.
It was some time before scientists realised that pressure waves corresponding to vertical seabed motions were propagated through the water column, and could be readily detected by a sensitive underwater pressure sensor or hydrophone, thus avoiding the need to place a geophone on the seabed. This development built on work done on the propogation of sound and shock waves through water as part of World War II research into surviving the shock of explosions when in the water and in defeating the various types of acoustic mines.
Data from single vertical geophone sensors or hydrophones is well show composite waveforms primarily dominated by the faster ‘p’ waves. It was only later in the evolution of marine seismics that scientists sought to obtain additional information from the shear wave energy. Shear waves produce predominantly horizontal motions of the seabed for near vertical incident seismic waves ( more or less the usual arrival direction) and do not propogate directly into the water column as fluids do not support shear wave propogation. Therefore to record and make use of this shear wave information the arrivals have to be detected at the seabed surface using a detector capable of recording three orthogonal components of motion. In order to distinguish the components, the orientation of the sensors with respect to the vertical must be known, and for some time this was achieved by gimballing the sensors in a weighted pendulum within a chamber containing an oil of moderate viscosity to ensure that the short term seismic frequency motion of the actual sensors followed that of the casing. The geophone housings for the gimballed sensors were of necessity rather large and cumbersome cylinders of around 230mm diameter and 300 mm height. I am not aware of any testing to check that the various gimbal systems actually did what was expected of them, particulary those used in the PUSS instruments that relied on static friction in the gimball bearings to communicate motion to the sensors in the absence of any liquid. The large cylindrical geophones were deployed either vertically or on their sides depending upon the requirements of the particular instrument rig, and were abandoned on the seabed at the end of the experiment by means of a cutter and release mechanism as part of the system that released the main instrument package from the anchor weight.
During one experiment in 1986 both orientations of geophone deployment were used in the same area and a radial pattern of shots to test for anisotropic propogation through old and new ocean crust, and it became apparent that there was little correlation between the shear wave signals seen on the different configurations, indeed they did not even display the same shear wave frequencies from the same shots. This led to a series of experiments by Tim Owen, initially in the lab at sea and later in more controlled conditions that showed the motion recorded by the sensors as horizontal motion of the ground was in fact rolling or wobbling of the housing on the seabed. Careful experiments as part of undergraduate physics projects by David MacCormack and others over several years investigated this phenomenon and led to the design of a completely new low profile three component geophone configuation without gimbals that relied on the fact that in almost all areas of the oceans the seabed is essentially flat due to sedimentation, or was within the 15 degree tolerance of the normal commercially available 4.5 Hz geophone elements. The low profile geophone developed in the lab continued in use academically and commercially for 25 years as the Carrack ‘phone.
The first experiments in marine seismics were based on land experience, and so used geophones placed on the seabed. It wasn’t until the post war revival of academic seismics that experiments were made to determine whether hydrophones would be as effective as geophones in recording seismic energy. As long as experiments were done with receivers placed on the bottom from stationary ships in shallow water there was no imperative to develop hydrophones, but as soon as the problems of organising seismic experiments around a single ship or in deep water were tackled, it became necessary to look at ways of deploying geophones from surface floating buoys. Hydrophones measure pressure changes, and in near surface deployments will respond to both seismic waves travelling through the water and to pressure changes due to changing depth of the hydrophone or pressure waves generated by waves on the water surface. A hyrophone simply suspended below a buoy will, under normal sea conditions, record nothing but noise generated by movement of the buoy and waves on the surface of the water driven by the wind, and it took considerable development and care in construction to come up with a satisfactory suspension for the hydrophone that would put the hydrophone below the surface noise depth and decouple it from the motion of the buoy due to wave action.
Hill Buoy hydrophone stabilisation
Early Maurice Hill piezoelectric Hydrophone – circa 1949
Valve hydrophone pre-amplifier for Hill experiments
This system was developed for the Hill sonobuoys, a later refinement for sonobuoys used a vertical wire with a weight at the bottom and a long horizontal arm made by attaching floats at intervals along the wire to achieve approximate neutral buoyancy, and with the hydrophone attached to a long float that was carefully adjusted using lead shot to be accurately neutrally buoyant in the sea area of the experiment to take account of the salinity and temperature. The slight motion through teh water due to wind drift probably helped to stream the suspension horizontally. Hydrophones were used as the sole detector until the department returned to putting recording packages incorporating geophones on the seabed ( PUSS 19XX), and by this time the hydrophone had become the standard sensor so these new instruments always included hydrophones as the primary sensor – fortunately in teh case of the PUSS as it turned out because the tape transport motors generated so much noise that the geophone signals were unusable until remedial action was taken.
This chronological account of seabed seismics covers a period of intense technological progress, where at each step the instrumentation developed was using the most up to date commercial technology, and where that technology was still barely adequate for the task in hand. At the birth of marine seismics in the late 1930s almost nothing was known about the problems of putting instrumentation in shallow water, let alone at oceanic depths, and almost nothing was known about the propagation of seismic waves from the seabed into the water column. All electronics depended on thermionic valves that needed voltages that were high for batteries, which themselves had very limited storage capacity compared to modern ones. Small D.C. electric motors were very inefficient because the best available magnets were weak and did not retain their magnetism for long. In the early days about the only method available for recording data was by using mirror galvanometers to direct a spot of light onto a moving strip of photographic paper, which had subsequently to be developed and fixed before the data could be checked or interpreted – the paper was bulky and needed to run at speed to give adequate time resolution, restricting its use for any seabed application. After the War 35mm film was substituted for the photographic paper, with some saving in space requirement, in the Woodhill galvanometer recorders developed for the Admiralty and used in some of our early seismic instruments, giving a maximum recording capacity of 12 channels including timing channels for a duration of XXXX minutes. The limited dynamic range of the analogue electronics and recording system, combined with unpredictable signal amplitudes usually meant that each sensor usually needed two signal channels associated with it to produce adequate dynamic range. The one technology that was reasonably advanced from the early days was wireless communication, particularly over ranges of several miles.
Woodhill 35 mm film galvanometer camera
All ‘active source’ seismics depends upon measuring the transit times of seismic waves and this in turn requires that we have the means to relate the shot time to the recorder time. There are many ways of achieving this, but in the early experiments stand-alone clocks did not have sufficient accuracy to run an independent remote timebase, and a number of systems were tried to solve the problem. The first experiments used sensors deployed on the seabed from the main ship, and a smaller launch to steam off and drop charges, with a radio signal generated in response to the explosion that was received aboard the ship and added as a timemark. This went out of favour after the Second World War because it was difficult to get two ships, so charges with timers and radio time break signals were floated off, until it was realised that this wasn’t perhaps the safest of techniques! Subsequently charges were floated off on very long lines and electrically fired from the ship. – not that much safer. The commercial introduction of the transistors in 1954 and the introduction of digital electronics thereafter meant that by the 1960s it was possible to make crystal oscillator clocks using up to date X-T cut crystals that would keep time accurately enough to run as a timebase in an autonomous recording system, allowing the sea bed recorders that were not connected to the surface by wire. By this time magnetic recording on tape had become a standard domestic process, although none of the commercially available recorders were directly suitable for seismic instrumentation and the lab devoted a considerable effort over the next 20 years modifying standard recorders and building special electronics, initally analogue system using frequency modulation. In 197X the laboratory purchased an early 8 bit computer, the ALPHA LSI with 16 Kbytes of core memory and a very expensive and complex interface to a half inch digital tape recorder that was used to digitise these analogue FM tapes on return to the laboratory. In 198X the lab developed the circuits necessary to record digital signals onto stereo analogue cassette recorders ( Sony professional Walkman using MFM encoding at 9600 bits/sec on 2 channels),using a data packet protocol that made the number of recorded channels independent of the number of parallel recording channels and did away with the need for a separate timing channel and duplicated channels at different gains. This system endured until computer technology produced true digital recording systems that were small enough and had a low enough power consumption for portable systems, with enough data capacity to allow useful recording durations – the first such being hard disks, which led to the development of the miniDobs – a truly digital system based entirely on computer technology. Development of this system was begun when the recording capacity of the available low powered hard disks was 20 Mbytes – barely enough for a short single channel experiment, but by the time development had finished adequate capacity for reasonable single channel experiments was commercially available at reasonable prices.
Initial seismic experiments were limited to shallow water depths and systems physically connected to the surface ship or a buoy as a truly autonomous pop-up system had not been developed in the UK. Early problems for tethered system related mostly to keeping the water out, it is one of the paradoxes of underwater instrumentation that it is easier to keep the water out of pressure vessels at significant depths where the water pressure forces well designed surfaces together and closes up seals, than it is in shallow water the pressure is enough to force water past poor seals without there being enough pressure to force surfaces together. In the early days much of the sealing was done with natural rubber bungs, for although O rings were invented in 19XX they did not become common engineering items for some years. The main constraint on autonomous pop-up systems in the early days of marine seismics was the lack of any convenient buoyancy that would survive depths of thousands of metres. The only real candidate in the 1940s and 50s was the use of a low density relatively incompressible fluid – and unfortunately the only readily available candidate was petrol, and it was necessary to use large volumes to get a useful lift – early American experiments used 30 gallons of petrol for a single instrument – not popular aboard ship! Pop-up instrumentation did not really become practical until the production of borosilicate glass hemispheres by Corning around 195X, but then they had the advantage that not only did they provide buoyancy, but had a useful internal volume into which the instrumentation could be fitted, provided that the overall package was still buoyant. Glass fulfils this role because, while it is weak in extension because it is liable to fracture, uniformly applied external pressure puts a sphere into compression and the ultimate strength of the material is realised. Bob Whitmarsh’s pop up seismometer is an early example of this type of seismic seabed instrumentation, although his early production spheres did not live up to the marketing hype. The initial problems were thought to be the result of having hemispheres with ground faces placed together, and fused hemispheres suitable only for buoyancy were produced. When the lab later returned to the design of pop-up instrumentation later, fused glass hemispheres were used for buoyancy and aluminium tubes for instrument housings, partly because a generation of shallow water sea bed seismometers ( the PUSSES ) had used aluminium tubes and the internal design could be replicated. The design of deep water pressure vessels for use with external pressure is more complicated than the same problem for internal pressure because there are two failure modes with external pressure, failure of the material, and shape collapse because the shape becomes unstable, as with a Coke can being crumpled, whereas with internal pressure only the material failure occurs because. When we started to build deep water seismic instruments we had a batch of high strength aluminium tube extruded for us and this lasted for several different instruments, but we had run out when we needed additional instruments using similar electronics, so we made a number with glass hemispheres as the instrument housing without any problems. When we came to design the miniDOBS the design brief was an instrument that we could build, store, transport and deploy in larger numbers and lower cost than before so the large aluminium tubes and multiple buoyancy spheres of the previous generation had to be abandoned in favour of a return to the form of our first pop-up seismometer, that of Bob Whitmarsh, with the whole system contained in a single glass sphere that doubled as instrument housing and buoyancy ( The glass spheres used were 17 inches O.D. and provided a nett buoyancy of 25Kg).
Pop-up seabed instruments work by having a heavy anchor that sinks the package to the bottom, and a release mechanism of some sort that releases the anchor and lets the rest of the instrument rise to the surface. The Whitmarsh pop-up instruments originally used a simple release based on a loop of magnesium alloy that gradually dissolved away in seawater and released the anchor – obviously a rather inexact timing system! The next iteration of seabed pop-ups using the big tubes and multiple buoyancy spheres used a release system developed by the Institute of Oceanogrphic that consisted of a substantial mechanical release latch activated by a pair of Pyro release elements that were based on an aluminium bar surrounded by a Thermite charge ignited by an electrical detonator. The detonators were activated by a 10KHz acoustic link housed in an aluminium tube. This release system itself required an additional 17 inch glass buoyancy sphere to bring it to the surface, so was a very costly part of the system both in money and payload terms.
When we came to design the miniDOBS it was obvious that we needed the whole instrument complete with its release system to be comparable in weight and cost to the existing release system in isolation. It was therefore necessary to design our own release actuator suitable for the lighter loads that would be involved, together with an acoustic system that could be housed within the single 17 inch sphere with the recording electronics. The resulting actuator was made from nylon and used a stainless burnwire that electrolysed away in a minute or two, and was more or less neutrally buoyant. We were fortunately able to purchase simple 10 Khz acoustic boards of circular form that fitted conveniently into the spheres.