“Statistical classification of diving behaviour in female rockhopper penguins from Staten Island, Argentina”

A RAYA REY; A SCHIAVINI

Ushuaia, Tierra del Fuego, Argentina

Conferencia; V International Penguin Conference; 2004

CADIC-CONICET

Time and depth recorders are capable of storing information over several weeks, yielding thousands of dives. Dives are usually characterized by variables strongly related. Multivariate methods provide both a useful tool to analyse huge amounts of data (dives), and to reduce the number of variables and human subjective bias in interpreting diving behaviour. 11819 dives of eight adult females of southern rockhopper penguins (Eudyptes chrysocome chrysocome) recorded during the chick rearing period in Tierra del Fuego were classified manually and statistically (principal component analysis, discriminant function analysis and cluster analysis). Using a combination of these methods all dives were classified into nine dive types. Each of them was characterized by a combination of the variables used to feed the multivariate analysis (like duration, bottom time, maximum depth, etc.). Visual inspection of the distribution of the dive types over the foraging trip context and time of the day, allowed to assign functionality to each dive type. The nine types were then reclassified into three hypothesized function groups: travelling, searching and foraging. The first one was characterized by shallow depths and short duration dives restricted at the beginning and end of a foraging trip. The exploration group was composed by dives of middle to intermediate depths, longer duration than the previous but with similar descent and ascent rates. The foraging group was similar to the second one but characterized by lower ascent rate. This characteristic was reported in other crustaceans’ eaters pursuing mobile patches of preys. Most of the dives belong to the foraging group and occurred often immediately after some searching dives type. The classification of dives could be used to estimate foraging effort and efficiency, by allowing calculation of the proportion of time spent performing dives associated with foraging and the relative number of feeding events recorded. Although such an index remain a coarse estimate of foraging effort and efficiency, it would provide further insight into key aspects of the foraging ecology of these animals.