I am running PAMGuard click detector on a two hydrophone array's recording to detect sperm whales clicks and calculate their bearing. The array is at a roughly 2300m depth, 40m above ocean's floor, recording 5 min every 10 min and the two hydrophones are 69 cm from one anothers. The click detector is roughly working, there is some over-detection but I'm fine with it as I am manually checking and annotating detections to separate false detections from first, second and third arrival of a click to the hydrophone (multipath), as well as separating between animals when there is more than one. The bearings calculated are sometimes lacking consistency however, with bearing jumping between two sets of values tens of degrees apart from one another. The graph below illustrated my problem. It shows the variations of bearing (in radians) calculated from the first arrival by PAMGuard over the time of an 'encounter' (a series of recordings with consistent click parameter leading to confidently think its the same animal over all the recordings). All recordings are problematic for this encounter but those between 4:10 and 4:15 or between 5:10 and 5:15 are the most striking.
The heading of the array is measured only every 30 seconds but it is rarely rotating tens of degrees in one direction or another.
I haven't modified any parameters related to bearing calculation before running PAMGuard, is there some and if yes where can I access them? If there is none where can I find information on how PAMGuard calculates its bearings?
