What methods are currently available to time synchronise acoustic recorders underwater automatically on a regular basis? I am looking to deploy a small 3-hydrophone triangular array of Soundtraps at roughly 2m spacing to locate baleen whales. However, the area is fairly remote so I am looking for a stand-alone solution to synchronise devices regularly while deployed. Deployment duration would be about 4-6 months.
Accurate localisation using time delay of arrival measurements requires three primary components;
- Accurate time delay measurements.
- Accurate positions of hydrophones.
- Knowledge of propagation conditions - in the most basic sense this would be a single sound speed but may be more complex if there are temperature, salinity, depth gradients that cause refraction of sound waves.
SoundTraps have an internal clock that drifts up to 2 seconds per day. Measuring the time delays based on the SoundTrap's clock is therefore impossible over all but the largest apertures because the drift will cause a huge error in the time delay measurements.
SoundTraps can also be set up to produce a pulse every second that can be used to synchronize connected devices, however, this is not suitable for all array configurations. In one second, the clock drift is 0.0231 milliseconds, or if we convert that to a distance based on a sound speed of 1500 m/s, this is equivalent to an error in hydrophone spacing of around 3.5cm. So for a 2 meter spacing and two SoundTraps, that means you have an error of sqrt(0.035^2+0.035^2)/2 = ~12 % in time delay measurements. That is actually quite a lot..The larger the array the less the percentage error and more accurate the localisations.
I would therefore recommend that synchronized single channel SoundTraps are only used on larger apertures of perhaps 10m or more. 4-channel SoundTraps are sample synchronised and do not have this issue.
The easiest way is to ping them with a sweep (to get good time resolution) from a fixed location with respect to the hydrophones. We normally place a weak high frequency source at one of the hydrophones and have that ping at some interval.
Edit: good time resolution is obtained by matched filtering