2.3 Duck and invertebrate surveys
The duck species studied here are ground nesting and distributed widely in the boreal zone: mallard, common teal (Anas crecca; hereafter teal), Eurasian wigeon (Mareca penelope ), northern pintail (Anas acuta ), northern shoveler (Spatula clypeata ), garganey (S. querquedula ) and tufted duck (Aythya fuligula ). All bred in Maaninka, but only the first two species bred at Evo. Duck pair and brood surveys were made in 2017 and 2018 with standard waterbird round count methods (Koskimies & Väisänen, 1991). In the round count the wetlands were surveyed by moving around the lake by a boat, stand up paddling board or by foot near the shoreline so that all the settled birds were detected with a high probability. Pair surveys were conducted in April and May right after the ice melt, when duck pairs occupy their breeding wetlands and are preparing for nesting. Ice melting sets an exact time frame for the duck surveys (Pöysä, 1996, 2019), making it possible to calibrate the phenology between different areas and years. Pairs and lone males were considered as pairs. Also groups of 2–4 males were used to estimate the number of pairs (i.e., 2–4 pairs). If the number of females surpassed the number of males at a wetland, the number of females was used instead. Brood surveys were conducted twice a year in the early June and July. The species, number and age of ducklings were recorded for each brood (Pirkola & Högmander, 1974). When studying brood habitat use, all brood observations were used for the analyses to determine the diverging habitat use of different age classes (i.e. some broods might occur twice in the analyses). When analysing brood production, every brood was identified based on their age and count, thus counting each brood only once.
Invertebrate trapping was conducted in the water bodies in June 2017 and 2018 during the first brood survey. All details of the trapping procedure were identical between the 46 studied water bodies. Free-swimming aquatic invertebrates were trapped with the activity trap described in Elmberg et al. (1992). Emerging insects were captured with emergence traps similar to those described by Danell and Sjöberg (1977). Three traps of both types were set per water bodies in different vegetation types for 48 hours. All the water bodies of each area were trapped during one week. The two invertebrate measures were combined to give a water body-level food abundance index (Holopainen et al.,2014) as an index of habitat quality (for more information about local invertebrate catch and species-specific duck-invertebrate associations, see Nummi et al., 2013; Nummi & Väänänen, 2001).