The Keck Endeavour Ocean Bottom Seismometer Network
The Keck Endeavour network consists of eight seismometers deployed in and around the Endeavour Segment in order to record earthquakes related to volcanic eruptions, tectonic activity, and the circulation of hydrothermal fluids. These instruments were deployed using different methods depending on the local geologic setting. Five seismometers were inserted into holes drilled into lava flows, two were placed in concrete blocks and deployed in sedimented areas, and one was placed into a caisson and buried in a muddy plain. Seven seismometers are short-period instruments that record relatively higher frequencies (1-50 Hz), and one is a broadband instrument that records lower frequency signals (0.003 - 25 Hz).
The short-period instruments consist of a sensor housed in a titanium rod measuring 40 cm in length. The sensor connects to a glass sphere that contains a data logger (four laptop hard drives and a motherboard) and a pair of lithium battery packs. The air in the logger spheres causes the unit to be positively buoyant, so they need to be connected to a heavy anchor (railroad ties) via steel cable to keep them on the seafloor. All the glass spheres are encased in a bright yellow "hard hat" made of heavy-duty plastic to make them easier to find on the seafloor and to provide the spheres with some extra protection.
The broadband instrument is housed in a titanium sphere and buried beneath the heavily sedimented seafloor through a long, arduous process. First, ROPOS uses its own weight to puncture a hole in the seafloor with a large piece of PVC pipe. Next, the ROV vacuums the sediment out of the hole it created and sets the sensor inside. Then, 80 pounds of glass beads are poured into the hole to completely surround the sensor. Finally, ROPOS connects the buried instrument to a metal frame that has the data logger sphere and three other spheres solely containing lithium battery packs.
The purpose for such a long installation procedure is to provide a better coupling of the instruments to the seafloor. Using an ROV to aid in the installation of the instruments is a novel way of deployment, and allows us the opportunity to more securely anchor the sensors to the seafloor. With a better coupling to the seafloor, we are able to remove the low frequency (1-2 Hz) tidal noise that is present in other data sets.
The network was first installed in 2003. In following years, seismometers were recovered and redeployed annually through 2005. However, after the 2005 deployment, the instruments were not recovered again until August 2007. This 23-month period was the longest these types of instrument had ever been underwater, and no one was certain if any of them would still be operational after being brought onto the ship. One source of worry was seawater entering the glass sphere and flooding it. Aside from ruining the hard drives, the saltwater could react with the lithium battery packs and produce a dangerous mixture of hydrogen gas. If a flooded logger rises to the surface, the decrease in pressure would cause the gas in the sphere to expand, possibly to the point of exploding. Obviously, this would destroy the data logger and possibly damage ROPOS. Since the yellow "hard hats" prevented us from directly inspecting the spheres for flooding, our only clue for avoiding this dangerous situation was to see if the cases had sunk to the seafloor. Any instruments we encountered in this position would be left alone.
Cautiously, we went about our business and overall had a very successful recovery. None of the data loggers were flooded, although we did find one lying on the seafloor from a previous deployment. Needless to say, we avoided it.
One problem we encountered on multiple occasions was the corrosion of copper clasps connecting the data loggers to their respective anchors. In saltwater, copper is a reactive metal that corrodes very easily. The reaction corroded these clasps so thoroughly that they would break and the data loggers would float in the water column attached solely by the much more delicate seismometer cable to the installed sensor. Luckily, there was no observable damage done to the seismometer sensor, and it did not seem to affect our recorded data. During recovery of the last instrument, one of these clasps broke before we could secure the sensor to ROPOS and the whole logger/sensor package free-floated to the surface. It rose more than 2000 m in about 40 minutes before it emerged on the ocean surface. However, we were still able to recover it from the waves using grappling hooks from the ship.
We also had issues recovering the broadband unit. The first time we tried, the data logger slipped out of ROPOS's grappling hook and the frame containing the data logger and batteries was left behind on the seafloor. So we dove again, and managed to bring the logger to the surface. However, a cable attaching ROPOS to the frame snapped and the whole package started to sink back to ocean depths! Thanks to the quick-thinking, hard-working crewmen onboard the Thompson, they used grapple hooks to secure the broadband system and bring it back onboard, safe and sound.
Other than fighting off a couple cases of bad luck, everything went pretty smoothly. We were able to recover over 175 gigabytes of data, and we didn't lose a single instrument or data logger. Next up is to process and analyze all our data, which is a far longer but (hopefully) more rewarding process.
