Fishing for drifts: detecting buoyancy changes of a top marine predator using a step-wise filtering method

摘要: In southern elephant seals (Mirounga leonina), fasting- and foraging-related fluctuations in body composition are reflected by buoyancy changes. Such changes can be monitored measuring the rate at which a seal drifts passively through water column, i.e. when all active swimming motion ceases. Here, we present an improved knowledge-based method for detecting from compressed abstracted dive profiles received telemetry. By step-wise filtering of data, developed algorithm identifies fragments dives that correspond to times animals drift. records 11 South Georgia, this identified 0.8-2.2% as drift dives, indicating large individual variation diving behaviour. The obtained time series exhibit that, beginning each migration, individuals were strongly negatively buoyant. Over following 75-150 days, peaked close or neutral buoyancy, indicative seal's foraging success. Independent verification with visually inspected detailed high-resolution data confirmed is capable reliably species using data. This also affirms convey geometric shape sufficient detail them identified. Further, it suggests method, could detected even old datasets information, conventional classification previously failed.