Show more...
Abstract:
This archive contains three tables: 1) Positional dataset based on GPS tracking data of Antarctic petrels (Thalassoica antarctica) from Svarthamaren breeding colony, Droning Maud Land. Data were collected during three consecutive breeding season (i.e. during the austral summer), from 2011-12 to 2013-14. Positions have been re-discretized (see methods below) at 30-min intervals. | 2) Table S1 from the Supplementary Information. | 3) Table S2 from the Supplementary Information.
Data associated to the following publication: Tarroux, A., Cherel, Y., Fauchald, P., Kato, A., Love, O., Ropert‐Coudert, Y., Spreen, G., Varpe, Ø., Weimerskirch, H., Yoccoz, N., Zahn, S. and Descamps, S. (2020), Foraging tactics in dynamic sea‐ice habitats affect individual state in a long‐ranging seabird. Funct Ecol. Accepted Author Manuscript. doi:10.1111/1365-2435.13632
Quality
Positional dataset - Methods: 1 - GPS coordinates were projected using package proj4 v.1.0-8 (Urbanek 2012) into Polar Stereographic coordinates with 70°S as standard parallel. All successive locations involving perfect stationarity were first removed, and all tracks were then filtered using stepwise recursive routines to exclude locations generating unrealistic movement rates (i.e., speeds > 40m/s; Tarroux et al. 2016). We examined each track twice, before and after filtering, to ensure that all obviously erroneous locations had been removed. The average proportion of relocations removed per track was 0.2% (range = [0; 2.3]). The filtered dataset obtained consisted in n = 132 071 individual GPS locations, representing 1 202 bird-days of tracking. 2 - The programmed time interval between successive GPS locations varied between 5 and 90 min (median = 10 min), and we therefore re-discretized (i.e. resampled at discrete intervals along the path) all tracks at 30-min intervals, thereby assuming linear movements and constant speed between successive relocations (Calenge 2006). Package adehabitatLT v.0.3.16 (Calenge 2006) was used to re-discretize the tracks. 3 - All locations situated 10 km or less from Svarthamaren breeding colony were excluded. The total number of locations obtained was n = 57 418. 4 - We applied hidden Markov models (HMMs; Boyd et al., 2014; Zucchini, MacDonald & Langrock, 2016) to identify the most likely sequence of behavioural states of an individual along its foraging track.See related publication for details.
Positional dataset - Variables description: || “birdID”: unique individual identifier || “track”: track number for each individual, independent of whether tracks were recorded successively or not (most individuals were tracked only once) || “timestamp”: format is %Y-%m-%d %H:%M:%S %z (‘z’ being the time zone) || “lon”: longitude || “lat”: latitude || “hmmState”: Behavioral state predicted by the HMMs. See related publication for details (Appendix 6 and Fig. S1). Levels: 1 (drifting), 2-3 (foraging), 4-5 (transiting).