Quantifying behaviours from acceleration: the case of the Caspian tern.

The Caspian tern, known as “skräntärna" på svenska, is the biggest tern in the world, inhabiting large lakes and ocean coasts in North America, Asia, Africa, Australia, including New Zealand and locally in northern and eastern Europe. The total world population size counts 420, 000 mature individuals and it seems to increase. However, in Sweden the tern species is listed as Least Concern and considered Vulnerable on the National Red List. It is caused by loss of breeding places, shallow bays and lakes suitable for fishing, and more recently by predation by invasive species like American Mink. The Swedish population consists of 600 pairs which is almost one third of the Baltic population and a major part of the total European population!

Fig. 1. Caspian terns (dancing?) at Norra Stenarna in June 2014. 
Picture taken by Ewa Karaszewska.

Studies on the behavior of the Caspian tern is the part of the bigger monitoring programme of the seabird species running at CAnMove at Lund University. The field work took place at the East Coast of Sweden south of Gävle. Together with Kozue Shiomi and Tom Evans, in cooperation with the project leader, ringer and local fisherman, we deployed tags on five seabird species: common gull, herring gull, lesser black- backed gull, greater black- backed gull and Caspian tern (Fig.1). The research station was situated in Fågelsundet, just a few kilometers by boat from the main colonies of seabirds at Stenarna (Fig.2) and Tågstumparna.

Fig. 2. Field work at Stenarna while ringing chicks. Visible hide. 
Photo taken by Kozue Shiomi.

CAnMove scientists in cooperation with the Swedish Ornithological Society, purchased modern GPS tags weighing only 2-3 % of the body mass and equipped with solar panel from scientists developing the equipment at Amsterdam University. Data and device management was possible via UvA BiTS (University of Amsterdam Bird Tracking System; UVA-BiTS). The tag enables to not only receive GPS positioning without recapture, but also quantify behavior using precise accelerometer and temperature sensor. We captured the birds using walk-in and spring traps located on the nests during incubation (Fig.3). 

  Fig. 3. Spring trap located on the gull’s nest (left). Tagged and marked Caspian tern (right). Picture taken by Ewa Karaszewska (left) and Kozue Shiomi (right).

The aim of my master project is to quantify behavior of Capian terns during breeding seasons, detect the main foraging areas and obtain the time budget. I used the telescope to identify tagged birds and the hand-set camera to record them during visit in accessible foraging areas, like coastal bay areas (Snöskär and Ledskär). The videos were used for semi-supervised training of the software that allowed quantification of behavioural classes from not annotated acceleration data. The task of gaining the videos was extremely difficult in the field, both caused by individual and seasonal changes in foraging trip location (often bird where absent in the place of observation) and limitation of devices (both detection, following and recording the bird was a challenge for one person). However, I was able to record almost all types of behaviours, including the successful fishing event.  The further analysis is in progress and soon we hope to present exciting videos with view on acceleration data.

But for now, enjoy the chicks of Caspian terns!  [video taken by Ewa Karaszewska]

 //Ewa Karaszewska