2.2 Artificial nest survival experiment with wildlife cameras
Artificial nest experiments were conducted in 2017 and 2018 to study nest predation rates at the water bodies. Nests were placed in sites where a dabbling duck hen could possibly lay a clutch, based on our own experience (nest site selection of boreal ducks is poorly studied, review by Holopainen et al., 2015; see also Väänänen et al., 2016). Some dabbling duck species nest at the shoreline, while others can place nests in the forest far from water bodies, so the artificial nest sites reflected this distribution. Each nest contained two farmed mallard eggs and some down from shot mallard females, mimicking the situation in the early stage of egg laying. Nests were constructed to resemble real ones: natural nest material from the nest surroundings was collected to form c. 20 cm wide nest cup and cover the eggs lightly. Eggs were not covered with down since ducks do not typically cover them before starting incubation. Nests were set under small trees or bushes, so they were hardly detectable from above. In open areas nests were established within tussocks.
Experimental nests were established in pairs around the water bodies: shoreline nests were placed less than five metres from shorelines and forest nests at least 70 metres from the shoreline nest (for more information, see Supplementary Appendix S2). In 2017 there were 46 nests at Evo and 42 at Maaninka, and in 2018 the numbers were 48 and 42, respectively (in total 178 nests). Density of the experimental nests was c. 1.2 nest/km2 at Evo and 0.9 nests/km2 at Maaninka. At Evo, every water bodies had only one nest pair a year, while at Maaninka there were fewer, but larger water bodies, and thus 2–3 nest pairs around the water bodies were established if possible. The same nest sites were used in both years to minimize the site effect.
Wildlife cameras (20 Uovision UV595-Full HD 12MP and four Niteforce Professional Trail Camera 12 MP) were set at artificial nests to identify predator nest visits and depredation time. Light-triggered passive wildlife cameras were ca. 1–1.5 m from nests, attached on trees or 1 m stakes. Cameras were active the whole seven day period responding to movement and were adjusted to take three pictures in a row, followed by a one-minute pause. The nest experiment was started during the pair survey (see 2.3 Duck and invertebrate surveys), the time when ducks initiate egg laying. One nest experiment round took seven days, and all the nests of a round were established and deconstructed on the same day between 9 am and 16 pm. Two rounds of nest experiments with different set of water bodies were carried out on each study area (i.e. two one-week experiments with 20–24 nests at the time). Nests were not visited by researchers during the seven-day exposure period.