4.1 Artificial nest survival
Interestingly, corresponding trade-off between food abundance and nest predation risk was not evident at the habitat-level, because ponds rich in food also have the lowest nest predation rates. The wildlife camera data show that the nests on the shores of seasonal, beaver or man-made ponds had higher survival than nests on the shoreline of permanent lakes. Because we tried to keep the nest cover constant between the experimental nests, this difference in survival rates probably arises from more heterogeneous shoreline habitats of ponds and/or the availability of other abundant food resources ponds offer for predators. It has been suggested that nest survival is a combination of large-scale environmental factors and local nest-site characteristics. Landscape productivity can affect general predator and prey abundance, but at the nest-site level vegetation and nest location might affect nest detectability and predator behaviour (Ringelman et al., 2018). For example, predator foraging in the landscape may be concentrated at habitat edges (Andrén, 1995), such as the interface between terrestrial and aquatic habitats. The occurrence of the edge effect may depend on the predator community and predator behaviour (Pasitschniak-Arts et al., 1998), and for instance whether productive wetlands attract and support high number of predators (Stephens et al., 2005).
In the Evo area, mammalian predators have been found to occur more often around beaver ponds than permanent lakes (Nummi et al., 2019b). Still, higher predator occurrence around the beaver ponds was not reflected in the nest predation results, indicating that indeed some habitat-related factors there are working in favour of higher nest survival. It is possible that variability in the shoreline creates circumstances under which predators probably are not able to form long-term search images, i.e. circumstances resembling those considered by Nams (1997) for prey aggregated in space or time. We suggest that because predators may use spatial memory to improve searching efficiency (Phillips et al., 2004), their search around permanent lakes is more regular and effective than around temporally unpredictable ponds. This underlines the potential importance of seasonally flooded ponds for breeding ducks, especially in agricultural areas.
Overall, nest predation risk was lower in forest compared to shoreline nests, indicating occurrence of the edge effect between terrestrial and aquatic ecotones. The pattern was not depending on the fields in the landscape but emerged in both forested and agricultural landscapes. Several currently threatened duck species in Finland are typically nesting near the shoreline (e.g. common pochard Aythya ferina , tufted duck), and may suffer from stronger nest predation rates than more flexible nesters (e.g. mallard and teal; e.g. Pöysä et al., 2019). Nesting in forest may be safer, but on the other hand, newly hatched ducklings will have to move a long way into water (Pöysä & Paasivaara, 2006).
Increased predator abundance and diversity are typical especially for fragmented landscapes (Andrén 1995; Pasitschniak-Arts & Messier, 1995). The pattern is particularly pronounced in agricultural landscapes, where there are already high numbers of predators, such as corvids (Andrén 1992; Roos 2002; Holopainen et al., 2020a). The results support these observations: nest predation risk was higher in the agricultural landscape, where high rates of corvid predation and richer predator communities were observed with similarly executed wildlife camera-artificial nest experiments (Holopainen et al. 2020a, 2020b). Wildlife camera studies conducted both with artificial nests (Holopainen et al. 2020b) and natural nests (Bell & Conover, 2023) have proven that after the initial depredation event, disturbed nests are often visited by multiple secondary predators. Multiple mammalian visits lead not only to an increased egg depredation rate, but also higher mortality risk for the incubating female. Indeed, hens often abandon (partially) depredated nests and even if incubation is continued, hatching success rate is low (Ackerman et al., 2003; Bell & Conover, 2023).
In Europe, the overall predator populations have increased during last decades threatening bird populations (Roos et al., 2018). In addition to native species, invasive alien predators such as raccoon dog have dispersed widely (Bonesi & Palazon, 2007; Kauhala & Kowalczyk, 2011). Raccoon dog nest predation can be destructive on islands (Dahl & Åhlén, 2018), but its role as mainland duck nest predator has remained unclear (Kauhala, 2004; Kauhala & Auniola, 2001; Nummi et al., 2019a; Sidorovich et al., 2008). Our nest predation results were based on detailed predation information derived from camera trapping. Corvids and raccoon dog were responsible for most of the nest depredation occurred at experimental nests mimicking the situation in the early stage of egg laying (Holopainen at al., 2020a). Without cameras, predator identification is uncertain as it relies on the remains of eggshells or other cues on the nest site (Larivière, 1999).
We recognise that artificial nests give an uncertain reflection of actual nest predation, and thus the intention in this study was not to evaluate actual predation rates but only to study the habitat-specific relative predation risk. Many important differences exist between real and artificial nests that decrease the correspondence and are thus recommended to consider whenever conducting artificial nest experiments (Butler & Rotella, 1998; Whelan et al., 1994; Wilson & Brittingham, 1998; Pärt & Wretenberg, 2002). Effort was put in to tackle the uncertainties: real mallard eggs were used and the species observed in the camera pictures are known predators of real duck nests (Opermanis et al., 2001; Pöysä et al., 1997), and therefore it is assumed that the observed species do not differ from the actual nest predator assemblage. As Anthony et al. (2006) showed with dusky Canada geese (Branta canadensis occidentalis ) artificial nests can be used to identify the potential nest predator species and that the predator species ratios can correspond those of the real nests. Our artificial nest density was low ensuring that observations were independent. The lacking hen problem was avoided by focusing only on the early egg-laying stage when females are not on their nests, so the set-up resembles the actual situation.
Compared to natural nests, the use of cameras may affect nest survival, typically by decreasing the predation rate (Richardson, et al, 2009); this is a potential shortcoming that could not be avoided. It is also acknowledged that this study design potentially emphasises the role of visual predators, such as corvids, as nests were not necessarily hidden as efficiently as a dabbling duck hen’s nest would be. High corvid predation rates may also be expected to occur at the early real nests, as the duck nest predation rate in North America during the early part of the breeding season was observed to positively relate to American crow (Corvus brachyrhynchos ) activity (Johnson et al., 1989).
The correspondence of the artificial nests with actual nest success cannot be assessed, but brood production rates in the study areas have been measured, which can be used as a rough estimate of predation pressure. While there are still uncertainties in this method, we emphasise that the problems underlined by the earlier studies have been considered and the differences between real and artificial nests were accordingly minimized; thus, we suggest that our data are suitable for detecting trends in predation rates in relation to habitat (Wilson & Brittingham, 1998).
Invertebrate food abundance
As expected, ponds (seasonal, beaver and man-made) were more invertebrate-rich habitats than permanent lakes, while contrary to the hypothesis, the percentage of field land around the wetlands did not influence invertebrate index. Selecting a pond instead of a lake as a breeding habitat would thus simultaneously minimize nest predation risk and maximize food availability in any landscape.
Habitat use of duck pairs was not associated with invertebrate food, whereas duck broods preferred habitats richer in food. The number of broods at the water bodies was only weakly dependent on the number of pairs, which can be a reflection of differing habitat requirements of pairs and broods (Holopainen et al., 2015) or high nest predation and brood mortality. Sjöberg et al. (2000) showed for mallards that all lakes used by pairs are not suitable for broods, the difference in lake use between pairs and broods being due to food limitation at the brood stage (Gunnarsson et al. , 2004). In boreal lakes food limitation can be intensified due to food competition between ducks and fish (Nummi et al., 2016). Income breeders, like teal, seem to avoid brood-stage food limitation by congregating in beaver ponds and seasonal ponds where invertebrate production is high and the habitat structure favourable for brood foraging (Nummi & Hahtola, 2008; Nummi & Holopainen 2014). In Evo it is known that teal brood production is following the flood dynamics created by the beaver and spring floods (Holopainen et al., 2014).
Interestingly, the results did not show that duck pairs or broods used ponds more than permanent lakes. This contradicts the earlier long-term results from the Evo area (Nummi & Holopainen, 2014). It is possible that ducks visit food-rich ponds for foraging in very short periods, reducing the ability to detect them there (Nummi et al., 2019c). Waterbird species may also differ in their ability to respond to environmental factors, such as habitat variability (Wiens, 1976; Nummi & Pöysä, 1997). For example, lapwings (Vanellus vanellus ) are known to nest in higher densities around flooded footdrains, and chicks forage on the wet mud around these wet features supporting invertebrate rich habitats (Eglington et al., 2008, 2010).
Conservation implications
Successful management of ducks would demand understanding of the relationship between habitat availability and predators (Drever et al., 2004). This study emphasizes the benefits of the availability of different water body types for breeding ducks. It showed that flooded and/or seasonal ponds might be especially good habitats. It seems that two important limiting factors of the breeding season – nest survival and amount of invertebrate food – are higher there than on permanent lakes.
Kubelka et al. (2018) showed that shorebirds have experienced a worldwide increase in nest predation over the past decades and that the pattern is especially pronounced in the high northern latitudes. 12 of the 19 duck species living in Finland are already classified as threatened to some degree by the Finnish red list (Lehikoinen et al., 2019), underlining the urgent need for conservation actions. These results indicate that while duck pair and brood densities are higher in an agricultural landscape, brood production seems to be higher in forested landscapes with lower nest predation rates. Thus, duck species nesting at eutrophic lakes in agricultural areas and preferring especially shorelines as nesting places, may suffer from high nest predation rates, which may contribute to the declining population trends (Lehikoinen et al., 2016; Pöysä & Linkola, 2021). It is suggested that the nest predation pressure around these lakes has increased due to the appearance of alien predators (Holopainen et al., 2021; Pöysä & Linkola, 2021) and the disappearance of protective umbrella species; for instance, loss of black-headed gullChroicocephalus ridibundus colonies removes a local protective “umbrella” of mobbing gulls, in a way that is thought to expose the nests of associated waterbird to greater predation threat (Pöysä et al., 2019). Controlling predators, especially alien species, would thus be an important conservation action to improve duck breeding success (Dahl & Åhlén, 2018; Garrettson & Rohwer, 2001; Jaatinen et al., 2022).
Seasonal pond ecosystems in the boreal biome remain poorly studied, even so that in Finland the habitat type does not have a conservation status evaluation done due to lacking information (Lammi et al. 2018). Nevertheless, seasonal pond habitats are commonly not protected. The loss of seasonal ponds has been dramatic in boreal biome due to drainage, destruction and water regulation (Colburn, 2004), also in Finland (Kuusisto et al., 1998). Furthermore, it is predicted that climate change will reduce the extent of snowmelt dependant spring flooding in the future (Veijalainen et al., 2010). In addition to wetland restoration and blocking up drains, the lack of flooded waters may be mitigated by favouring beavers (Hood & Bayley, 2008; Nummi & Holopainen, 2020) or man-made wetlands (Danell & Sjöberg, 1982; Eglington et al., 2008; Čehovská et al., 2022).