| Summary: | Intra-specific and intra-population variation inmovement tactics have been observed in many species,sometimes in association with alternative foraging techniquesor large-scale habitat selection. However, whetheranimals adjust their small-scale habitat selection accordingto their large-scale tactics has rarely been studied. Thisstudy identified two large-scale movement tactics in ringedseals (Phoca hispida) during their non-breeding, post-moultingperiod. First-passage times (FPT) were used toexplore these large-scale patterns. Subsequently, habitatselection was quantified by modelling the FPTs as afunction of habitat attributes using Cox proportional hazardsmodels. Some seals moved far offshore into areaspreferentially containing 40–80% ice coverage, while otherindividuals spread along the coasts of Svalbard concentratingtheir time near glacier fronts. Both tactics resultedin ringed seals being in highly productive areas where theyhad access to ice-platforms to rest. When offshore, habitatselection was influenced mainly by sea ice concentrationand season. Late in the season (autumn), increased risk ofleaving an area was identified, even when ice conditionswere still favourable, reflecting their need to return to overwintering/breeding areas before the fjords of the archipelagofreeze. For ringed seals that remained inshore, habitatuse intensities were influenced mainly by the distance toglacier fronts and season. These animals were already closeto their over-wintering habitat and hence their risk ofleaving an area decreased as winter approached. This studyof ringed seals habitat selection reveals how they fulfil theirbiological requirements in this dynamic, heterogeneoushabitat. Individuals within the same population employedtwo distinct large-scale movement tactics, adjusting theirdecisions for small-scale habitat selection accordingly.This flexibility in ringed seal spatial ecology during summerand fall is expected to result in increased populationviability in this high Arctic environment.Methods and data filteringA total of 22 animals, 11 caught in 2002 and 11 caught in 2003, were equipped with satellite-relayed data loggers (SRDLs). These SRDLs collect and relay information on movements (geographical locations), time spent dry (haulout behaviour) and aspects of diving behaviour (ascent rate, descent rate, time at depth, etc.) of marine mammals. In addition, they are equipped with a fast-response temperature sensor (micro-BetaCHIP probe; Betatherm, Ireland), with an accuracy of 0.1°C and a resolution of < 0.01°C. SRDLs were programmed to send data whenever possible, with no duty cycling.Location data obtained from Argos satellites were filtered using an algorithm based on swimming speed, distance between successive locations and turning angles. The filter removed all data points for which the Argos location process failed. It also removed all locations requiring swimming speeds higher than 2 m/s, unless the data point in question was located less than 5,000 m from the previous location. The latter condition enabled the retention of good-quality locations for which high swimming speeds resulted from locations being taken very close in time. Additionally, the filter excluded sudden deviations away from the general path’s track if the off-track spike extended greater than 2,500 or 5,000 m, with angles smaller than 15° and 25° between consecutive locations, respectively. Distance and angle limits were chosen empirically, based on the measurement of the angles and lengths of the most conspicuous, abrupt deviations from the principal path of the tracks, which are unlikely to correspond to real locations.
|
|---|