Population Decline in an Avian Aerial Insectivore (Tachycineta bicolor)

With an estimated 21-52% of the world’s biodiversity at risk of extinction, we may currently be experiencing Earth’s sixth mass extinction event (Barnosky et al. 2011; Dirzo et al. 2014). Biodiversity loss is a serious issue. First, biodiversity is intrinsically valuable, displaying the fantastic array of possibilities that evolution has generated over the millennia (Leopold 1949; McCauley 2006). Second, we enjoy spending time in nature, and doing so is good for our health and wellbeing (reviewed in Hartig et al. 2014). Finally, biodiversity is important to ecosystem stability and function. Ecosystems with high biodiversity produce more goods and recycle more waste than ecosystems with low biodiversity, and these benefits only increase with time (Reich et al. 2012). Biodiverse ecosystems are more stable, and less prone to cascading extinctions because they have many different connections between species (e.g., species A eats species B, C, and D, not just species B) (Dunne et al. 2002). Thus, biodiversity loss threatens ecosystem function and ecosystem stability, as well as our personal wellbeing. Regional biodiversity loss starts with the loss of a single species. Across a region, not all habitat is suitable for any one species, resulting in collections of smaller populations distributed across patches of suitable habitat, which are connected by dispersal of individuals between populations, forming locally inter-connected metapopulations (reviewed in Hanski 1997). Within a population, four demographic processes influence population size: birth, death, immigration, and emigration. In high-quality habitats, birth rates often surpass death rates, so there may be more offspring than there are available territories, forcing young individuals to emigrate elsewhere. These populations with net outflux of individuals are often referred to as source populations. In contrast, in lower quality habitat patches, there may be lower birth rates and/or higher death rates, forming sink populations. For a sink population to persist over time, it must 2 be sustained by immigrants from other populations (Brawn & Robinson 1996). Determining whether a population is a source or a sink can be difficult but is crucial to understanding how a single population fits into the surrounding metapopulation (Brawn & Robinson 1996). If overall patch quality declines, source populations may not be able to sustain sink populations or may become sink populations themselves. Across the world, broad-scale environmental change is influencing population dynamics of numerous species. Habitat loss, typically caused by increasing human land use and development, is the single largest threat facing species at risk worldwide (Sala et al. 2000), affecting at least an estimated 33% of all threatened species (Gurevitch & Padilla 2004). As we develop large swaths of land, patches of suitable habitat get ever smaller and farther apart causing overall biodiversity loss and preventing ecosystems from functioning effectively (Fischer & Lindenmayer 2007; Haddad et al. 2015). For long-lived species in particular, extinction and biodiversity loss can occur many years after the initial habitat loss (Tilman et al. 1994; Krauss et al. 2010; Haddad et al. 2015). We will likely reap the consequences of our habitat destruction for decades to come, regardless of current conservation efforts. Another way we are influencing populations around the world is through climate change. Climate change is expected to have the largest impact on arctic, alpine, and boreal habitats, and to influence other habitat types to a lesser degree (Sala et al. 2000), but tropical species may also be strongly impacted as they generally have relatively narrow ranges of thermal tolerance (Deutsch et al. 2008). Worldwide, species are responding to climate change by shifting their ranges poleward by an average of 0.61km yearly to track suitable habitat (Sala et al. 2000; Parmesan & Yohe 2003). However, many species are not able to shift their range because of ecological constraints (e.g., species with limited dispersal abilities, or species already at high latitudes or elevations). In fact, species are predicted to lose an average of 21-26% of their ranges by 2050 (Jetz et al. 2007). Additionally, many species are responding to climate change by breeding earlier as spring temperatures warm (Dunn & Winkler 1999; Sala et al. 2000). Unfortunately, breeding earlier is not advantageous if prey species have not shifted to the same degree, causing mismatches between offspring demand for food and peak food availability. Mismatches are common because there is substantial variation in the magnitude of different species’ phenological responses to climate change (reviewed in Walther et al. 2002), which can cause interacting species to become temporally disconnected. Food availability mismatches caused by climate change have been demonstrated in a number of species and linked to population declines (e.g., Gaston et al. 2009, McKinnon et al. 2012, Visser et al. 2012). As the global climate continues to change, its effects on species loss will become ever more apparent. Avian aerial insectivore decline One taxonomic group facing strong threats of extinction is birds (Butchart et al. 2004; Barnosky et al. 2011). In fact, current bird extinction rates have already exceeded the prehistoric extinction rates estimated during the previous five mass extinctions (Barnosky et al. 2011), making birds of particular concern for conservation efforts. Because birds are incredibly diverse and experience an equally diverse array of threats, finding one conservation strategy to address declines across all birds is an impossible task. Among birds, aerial insectivores, a guild of birds that forage on flying insects, are declining particularly quickly (Nebel et al. 2010; Inger et al. 2015). Avian aerial insectivores are a taxonomically diverse group, including swifts, nightjars, swallows, and flycatchers and encompassing more than thirty species in North America alone. Despite their distant phylogenetic relations, avian aerial insectivores as a group have been in marked decline since the 1980s, and exhibit particularly strong declines in northeastern North America (Nebel et al. 2010, Smith et al. 2015, but see Michel et al. 2016). Alternative hypotheses to explain aerial insectivore population decline The taxonomic diversity of avian aerial insectivores suggests that something about their shared feeding strategy makes them susceptible to large-scale environmental challenges (Nebel et al. 2010). A wide array of hypotheses has been proposed to explain population declines in this guild of birds, but a few have become prominent in the literature and are supported by some empirical evidence. Testing these non-exclusive hypotheses, which I describe below, is the focus of this thesis. H1: Aerial insectivore declines are caused by impacts of land use on insect abundance. Many avian aerial insectivores forage in large open spaces, many of which are disappearing or being degraded. Low-intensity agricultural fields are increasingly being converted to high-intensity agriculture, allowed to regenerate to forest, or urbanized, none of which provide suitable habitat for aerial insectivores (reviewed in Rey Benayas 2007). Increasing forest cover likely limits aerial insectivores’ line of sight and foraging efficiently, consequently preventing them from inhabiting reforested areas (Purves 2015). Aerial insectivores are uncommon in urbanized areas, suggesting that insect food or other important resources are unavailable in urban areas and these areas are avoided (English et al. 2017). Insect availability in agricultural areas depends largely on what type of farming activity is taking place on the landscape, with cattle pasture and hedgerows between crops exhibiting the highest insect abundances (Moller 2001; Evans et al. 2007; Grüebler et al. 2007; Rioux Paquette et al. 2013). As agricultural land use has shifted away from low-intensity cattle farming towards high-intensity cropland, birds that forage or breed in grasslands have declined substantially (Murphy & Moore 2003). While these grassland species are not exclusively aerial insectivores, one common avian aerial insectivore, the barn swallow (Hirundo rustica), is consistently more abundant in years following high insect abundance when agriculture is less intense (i.e., more cattle pasture, less total agricultural land, less pesticide and fertilizer use) (Moller 2001; Benton et al. 2002; Evans et al. 2007). Shifting land use patterns may affect aerial insectivores on both the breeding and overwintering grounds. Overall, the shift away from small-scale farms has lowered insect abundance and may be contributing to declines in avian aerial insectivore populations. H2: Aerial insectivore declines are caused by global climate change and its impacts on insect abundance and availability. Global climate change may challenge avian aerial insectivores by reducing food availability. Climate change affects temperature, wind, and precipitation patterns, all of which correlate with avian aerial insectivore population sizes (Irons et al. 2017; McArthur et al. 2017; Weegman et al. 2017). In northeastern North America where declines are strongest, climate change is predicted to cause increased precipitation and more variable temperatures with both more heat waves and cold snaps (Kunkel et al. 2013). Insect prey decrease with decreasing temperature (Cucco & Malacarne 1996; Grüebler et al. 2007), likely because the ectothermic insects are less active in the short-term and grow and reproduce less over the long-term. Notably, many aerial insectivores almost exclusively forage on the wing; that is, they forage while flying, catching insects that are also flying, so even in the absence of decreases in overall insect abundance, periods of inclement weather (windy, rainy, cold) can effectively reduce food availability to zero for these specialized foragers. Cold snaps may reduce adults’ ability to feed nestlings, or even cause nest abandonment if adults prioritize their own survival over their nestlings’ (McCarty & Winkler 1999a; Winkler et al. 2013; Ouyang et al. 2015). Similarly, poor foraging success overwinter could hinder adult survival. H3: Aerial insectivore declines are caused by global climate change induced mismatches between the timing of high-quality food availability and peak nutritional demand by nestlings. Nestlings may be particularly sensitive to climate change because of mismatch between demand and supply of insects. For one avian aerial insectivore, the tree swallow (Tachycineta bicolor), egg-laying shifted five days earlier from 1959 to 1991 (Dunn & Winkler 1999). Some evidence suggests that insect abundance increases over the course of the breeding season, which has been interpreted as evidence that mismatches between insect abundance and offspring nutritional demand are unlikely in these populations (Dunn et al. 2011). However, nestling growth increases more with higher quality food with more fatty acids than simply with more food overall (Twining et al. 2016). Aquatic insects contain more high-quality fatty acids than do terrestrial insects (Twining et al. 2016). Aquatic insects peak in abundance earlier than terrestrial insects (Nakano & Murakami 2001), and, in some populations, do exhibit mismatch with nestling demand for high-quality food (Twining et al. 2016). Therefore, climate change may influence avian aerial insectivore population dynamics by creating a mismatch between nestling demand and the emergence of high-quality food. Tree swallows as a model species Although the factors described above might be affecting all avian aerial insectivores, for most species we lack the long-term, detailed data to be able to evaluate even the basic demographic processes responsible for population decline. Tree swallows are an ideal model species for studying avian aerial insectivore decline. Naturally nesting in cavities excavated by other species, tree swallows take readily to artificial nest boxes, making them easy to study (Robertson et al. 1992; Jones 2003). Accordingly, researchers across North America have studied box-nesting populations since the mid-1960s, collecting a wealth of data along the way (Jones 2003; Shutler et al. 2012). Like other avian aerial insectivores, tree swallows show more severe population decline in northeastern North America (Nebel et al. 2010, Shutler et al. 2012, Smith et al. 2015, Michel et al. 2016). Additionally, tree swallow population declines began in the 1980s, which is congruent with patterns of decline in other species (Nebel et al. 2010, Shutler et al. 2012, Smith et al. 2015, but see Michel et al. 2016). Because tree swallows exhibit similar spatial and temporal trends in population declines as other avian aerial insectivores, tree swallows are a suitable model organism to study avian aerial insectivore decline. Tree swallows are an avian aerial insectivore from the family Hirundinidae. They eat aerial insects, predominantly and preferentially 3-5mm Diptera, although they also eat large amounts of Hemiptera and Odonata (Quinney & Ankney 1985; McCarty & Winkler 1999a). Tree swallows primarily rely on insects foraged during the breeding season to produce eggs (Winkler & Allen 1995; Nooker et al. 2005; Ardia et al. 2006), incubate, and provision nestlings (Winkler & Allen 1995; Nooker et al. 2005). Although they eat almost exclusively aerial insects during the breeding season, during winter, in times of insect shortage, tree swallows will supplement their diet with lipid-rich fruits and vegetation (Robertson et al. 1992; Piland & Winkler 2015). Tree swallows migrate south to the southern U.S. and Mexico, and overwinter roosting in flocks of hundreds of thousands of birds in coastal habitat along the Gulf of Mexico (Robertson et al. 1992; Winkler 2006). Migrating and wintering birds often roost in agricultural habitat, most notably sugar cane fields (Robertson et al. 1992). First-time breeders usually return to breed within 10km of their natal population (Winkler et al. 2005). Adults typically also return within 10km from their previous breeding sites (Winkler et al. 2004) but have been observed dispersing farther following unsuccessful breeding (Winkler et al. 2004, Lagrange et al. 2017, but see Shutler and Clark 2003). Tree swallows are one of only two North American species that display delayed plumage maturation in females, but not males: one-year-old females have brown body feathers in their plumage, and two-year-old or older females have full adult, iridescent blue-green plumage (Hussell 1983). The delayed plumage maturation in females has allowed comparison of one-year-old females with older females, revealing that older females breed earlier in the season and have higher reproductive success (Stutchbury & Robertson 1988; Bentz & Siefferman 2013). One-year-old females are also subordinate and less likely to occupy a territory if territories are limiting, resulting in “floater” populations of young females (Stutchbury & Robertson 1985, 1987). Territories can often be limiting as tree swallows are secondary cavity nesters, relying on cavities excavated by other species in dead trees in beaver ponds and other open spaces (Rendell & Robertson 1989).

Thesis objectives

The overarching aim of this thesis is to determine why tree swallows, and avian aerial insectivores more generally, are declining. To address this aim, I analyzed long-term data (1975- 2017) from a box-nesting population of tree swallows at the Queen’s University Biological Station. Each year, an average of 168 boxes (range 77-264) were monitored. Researchers have tracked nesting success in each box, measuring when egg laying starts, the total clutch size, when incubation starts, when and how many eggs hatch, and when and how many nestlings survive to leave the nest (i.e., fledge). Additionally, nestlings and adults were measured for body size and body mass and banded with uniquely-numbered Canadian Wildlife Service leg bands, allowing individuals to be tracked across years. In total, from 1975-2017, 5,506 nests, 18,366 nestlings, and 6,355 unique adults have been monitored. As the purpose of this data collection was originally to answer fundamental biological questions, many nests were experimentally manipulated. While experimentation may have influenced population dynamics in our nest boxes, it is not the cause of regional decline, so I have focused my analyses on unmanipulated nests. The intensive and long-term nature of our data (spanning both before and during population declines) gives us an unprecedented opportunity to investigate why tree swallow populations are declining. In chapter 2, I assess the relative importance of each of the demographic transitions between life stages (egg, nestling, fledgling, one-year-old, two+-year-old) to population growth rate using a life stage simulation analysis. For instance, does a small change in hatch rate create a larger change in population growth than a small change in adult overwinter survival? Answers to questions like this tell us which demographic transitions have the greatest impact on population growth. Such an analysis can inform conservation efforts, as efforts will be most effective when focused on these influential transitions. Similarly, environmental threats have bigger impacts on population size when they affect one of these important transitions. Building on chapter 2, in chapter 3, I asked whether the demographic transitions that I determined to be most important to population growth, overwinter survival and fledging success, have changed concurrently with the overall population decline. Specifically, I tested whether overwintering and migratory habitat loss, local weather post-fledging and/or global climate indices explain trends in survival overwinter and whether local weather patterns explain trends in nestling fledging success. Given the importance of nestling growth in determining fledging success and juvenile survival (Tinbergen & Boerlijst 1990; Michaud & Leonard 2000; Cleasby et al. 2010; Maness & Anderson 2013), in chapter 4, I asked whether nestling body mass close to fledging has changed over time. Using detailed growth curves and parental provisioning data collected during the 2017 breeding season, I also investigated the effect of local weather conditions (temperature, precipitation, and wind speed) and parental effort on nestling growth. Finally, I assessed whether weather conditions during nestling development have deteriorated over the study period. Ultimately, my research suggests that climate change, particularly increasing spring rainfall, may be responsible for tree swallow population declines by decreasing fledging success and juvenile survival. Understanding the role of climate change in these declines is crucial for acknowledging both which conservation management strategies are likely to be most effective and the unfortunate reality that, without stemming climate change, these strategies are likely to only be stop-gap measures