Migratory birds are increasingly exposed to urban environments as anthropogenic activities lead to altered land use practices. Understanding the impact of this environmental change on migratory birds is essential. Over 10% of migratory birds are currently threatened worldwide, and this percentage will likely increase due to habitat loss and climate change (Runge et al., 2015; Zurell et al., 2018). However, it can be difficult and costly to study migratory populations, as they inhabit different global regions and ecosystem types throughout the year. For migratory species that persist in urban environments, existing research demonstrates that living in urban areas can affect avian ecology and physiology through changes in habitat use, diet, exposure to pollution, and stress levels (Ruiz et al., 2002; Meillère et al., 2016; Murray et al., 2019). In addition, urbanization appears to contribute to changes in long-distance movements. Over the past decades, there has been an increasing prevalence of partially migratory species (i.e., species in which some populations undergo seasonal long-distance flights to and from specific regions while others do not). In many cases, the non-migratory individuals of these species exhibit year-round residency in urban areas, likely because of consistent resource availability and more moderate microclimates (Wiener and Tuljapurkar, 1994; Møller et al., 2014; Bonnet-Lebrun et al., 2020). One study found that living in urban areas increased the likelihood of residency in 9 of 12 partially migratory species (Bonnet-Lebrun et al., 2020). Another reported that shifts in the timing of food availability and breeding due to climate change favored shorter migrations and year-long residency in South German Blackcaps (Sylvia atricapilla) (Pulido and Berthold, 2010; Soriano-Redondo et al., 2020).

While some evidence indicates that resident birds have a larger average mass than their migratory counterparts, the impact of year-long urban residency on the health and physiology of partially migratory birds is not well-studied (Pérez-Tris and Tellería, 2002). Gut microbiome analysis offers a new non-invasive tool for studying these outcomes because it plays a critical role in host immune development, hormone regulation, and digestion in a variety of animal species (West et al., 2019; Mueller et al., 2020). The avian gut microbiome is strongly influenced by environmental factors such as habitat and diet (Hird et al., 2014; Waite and Taylor, 2015; Xie et al., 2016; Grond et al., 2017; Wu et al., 2018; Cao et al., 2020); therefore, changes in microbiome composition and/or diversity can have important consequences for avian health and fitness.

The effects of migration and urbanization on bird microbiomes has only been studied in a few populations. However, both migration and urbanization appear to affect the microbiome. Because migratory birds experience marked shifts in their environment, they are also more likely to exhibit shifts in the microbiome over time (Liu and Swanson, 2014). For instance, certain species of migratory birds change their diet preceding and during their migrations, which could result in a shift in the microbiome during migration (McLandress and Raveling, 1981). In migratory swan geese (Anser cygnoides), microbial communities shift between their breeding grounds and wintering area, with a core group of shared microbial taxa at both locations (Wu et al., 2018). Additionally, migratory birds often exhibit heightened stress responses during their migration due to the demanding physical toll of migration, and heightened stress has been linked to temporary changes in the microbiome in some animals (Liu and Swanson, 2014). The extent to which these dynamics are beneficial to hosts or induce negative health consequences is unclear (Risely et al., 2018). In contrast, urbanization appears to result in lower microbiome diversity in many current studies. Urban American white ibises (Eudocimus albus), house sparrows (Passer domesticus), and herring gulls (Larus argentatus) all exhibit lower microbial diversity (Fuirst et al., 2018; Teyssier et al., 2018; Murray et al., 2020), potentially due to the lower environmental diversity and smaller home ranges. In several bird species, the presence of Firmicutes, a key microbial taxon for dietary carbohydrate degradation, is negatively correlated with increased urban land cover (Drovetski et al., 2018; Wu et al., 2018; Murray et al., 2020). Therefore, year-long urban resident individuals of partially migratory species are likely to have distinct microbiomes compared to migratory individuals.

To begin to test this broad hypothesis, we assessed microbiome composition in Canada geese (Branta canadensis) in and around Chicago, Illinois, United States. The Canada goose is a large anatid waterfowl native to most of northern North America (Smith et al., 1999). Some populations of Canada geese migrate in fall and spring to their respective wintering and summering grounds, while others remain in one home range year-round (Conover, 1991). As a result, Canada geese have become common in urban areas and tend to be year-round residents, primarily feeding on lawn grasses, although migratory flocks can also be found in urban areas (Conover, 1991). The physiological differences between migratory and resident populations have been seldom studied. Although urban habitats may provide some benefits to geese, such as protection from predators and human hunting (Balkcom, 2010), urban-associated shifts in diet and movements might shift microbiome composition and diversity, affecting their health.

In this study, we assessed the microbiome composition of fecal samples from GPS-tracked individual Canada Geese categorized as either resident or migratory but occupying the same urban areas in Chicago, Illinois, United States. Based on the existing literature, we had two main predictions. First, we predicted that migrants would exhibit greater microbial diversity and inter-individual variation in microbiome composition compared to urban residents because they are exposed to more diverse habitats. Second, we predicted that migrants would have higher relative abundances of beneficial bacteria such as Firmicutes relative to residents because of their increased energetic demands. Given the current paucity of data in this area, our results will provide an important foundation for understanding the effects of year-long residency on avian microbiomes and, ultimately, physiology and health.

Contrary to our predictions, microbial alpha diversity was higher in resident geese across all three metrics used (Figure 2, Table 1 and Supplementary Table F). Also counter to our predictions, inter-individual variance (beta diversity) was also significantly higher in residents than in migrants (unweighted UniFrac: F = 17.57, p = 0.00021, variance = 0.0075; weighted UniFrac: F = 8.05, p = 0.0080, variance = 0.014; Figure 3). Migrant status significantly explained overall community composition (PERMANOVA, unweighted UniFrac: r² = 0.057, p < 0.05; weighted UniFrac: r² = 0.081, p < 0.05) (Figure 4 and Table 2). Community composition did not significantly differ by sampling site (Table 2). Geographical distance and microbiome dissimilarity were not significantly correlated (Table 3). To ensure that our results were not a result of our unbalanced sample size between migrant and resident geese, we reran each of these analyses 40 times using a random subset of nine residents and nine migrants. All of our results remained consistent when we subsampled, except for the weighted UniFrac community composition (Supplementary Tables G–I).

Several taxa exhibited significantly higher relative abundances (q < 0.05) in migrant geese compared to resident geese (Table 4). No taxa had significantly higher relative abundances in residents relative to migrants. Supporting our predictions, at the phylum level, there was a significantly higher average relative abundance of Firmicutes in migrants (approximately 81.36%) relative to residents (53.72%) (Figure 5). Firmicutes was the highest abundant phylum in 27 out of 33 samples (Figure 6). Within the phylum Firmicutes, four genera were significantly more relatively abundant in migrants than in residents: Terrisporobacter, Turicibacter, Cellulosilyticum, and Epulopiscium. Ten ASVs had significantly higher relative abundances in migrants compared to residents, eight of which were in the order Clostridiales (Table 4).

In this study, we tested the hypothesis that year-long urban resident Canada geese have a distinct microbiome compared to migrants. While our overall hypothesis was supported, contrary to our predictions, we found that migrant geese had less overall diversity and less within-group variation compared to residents. However, in support of our predictions, migrants had a higher relative abundance of Firmicutes, a bacterial phylum that is believed to be important for digestion and metabolism in wild birds (Grond et al., 2018) and has been associated with starch and cellulose breakdown in humans (Flint et al., 2012).

Given the wide range of ecological and physiological factors that differ between migrant and year-long urban resident individuals of partially migratory species, it is perhaps not surprising that we detected a difference in the gut microbiome composition of migrant and resident Canada geese in the Chicago area. Interestingly, however, the differences in overall composition were only detectable using an unweighted UniFrac method. This pattern suggests that the presence/absence of rarer microbial taxa differ more between groups than the relative abundances of the more abundant taxa and that migrant and resident geese share a substantial portion of their microbiome. This finding is similar to that reported for migrating migratory swan geese sampled before and after migration (Wu et al., 2018) and signals the potential presence of a core microbiome that is resilient to environmental change.

While the lower microbial diversity in migrants contradicted our original prediction, there are at least two non-exclusive potential explanations for this pattern. First, we expected migrant Canada geese to exhibit higher microbial diversity than residents because they are exposed to multiple habitats with potentially distinct dietary and environmental microbial inputs during migration. The rapid digestive rates and frequent foraging of geese (i.e., retention rate of approximately 2 h for other goose species; Prop and Vulink, 1992) would facilitate goose uptake of environmental microbes (Drovetski et al., 2018). However, it is possible that this uptake does not occur. If geese incorporate environmental microbiomes readily into their own we would expect the microbiome of migrant geese to converge with those of resident geese by the time we sampled them in an urban environment up to a week after arrival, but migrant microbiome composition was significantly different than residents. Similar patterns of reduced microbial diversity in migrant red-necked stints (Calidris ruficollis) (Risely et al., 2017) and curlew sandpipers (Calidris ferruginea) (Risely et al., 2018) compared to resident conspecifics have also been reported. It is possible that migrant microbiome resistance to invasion by environmental bacteria could be an adaptive mechanism that prevents pathogenic bacteria from being incorporated into the microbiome as migrants are exposed to so many new environments in a short time period (Risely et al., 2017).

Additionally, the resident geese in this study lived in urban areas year-round, which likely affects goose microbiome diversity and composition. Multiple studies have shown that increased urbanization alters the bird gut microbiome (Teyssier et al., 2018; Murray et al., 2020). While there are a variety of environmental factors that differ between urban and rural habitats, it is possible that diet may play an important role in driving observed patterns. Canada geese are known to actively seek out and anticipate food handouts from humans in urban areas (Conover, 1991). Resident urban geese may be able to exploit these artificial resources better than migrants, as they spend more time in the same place and have more time to learn where handouts occur. Geese that seek out human food sources are known to also eat plant matter (Conover, 1991), and so geese that consume anthropogenic food may have a more diverse diet than geese that purely rely on plants, and this greater diversity in diet could lead to greater diversity in gut microbes. Whether resident urban Canada geese actually receive more food from humans (or have a more diverse diet) than their migratory counterparts is unknown. However, similar to our results, polar bears (Ursus maritimus) that forage on land, and potentially include anthropogenic food waste in their diets, exhibit higher gut microbiome alpha diversity and a higher abundance of Firmicutes bacteria than bears foraging on sea ice (Watson et al., 2019). Baboons that feed in closer proximity to humans also have higher gut microbiome diversity (Barelli et al., 2020). In coyotes (Canis latrans), urbanization is correlated with greater anthropogenic food consumption, higher gut microbiome diversity, and unhealthy body conditions (Sugden et al., 2020). To better assess these relationships in geese, future studies should incorporate a quantitative assessment of diet and other environmental factors in urban and rural habitats.

With respect to variation in specific microbial taxa, the relative abundance of bacteria in the phylum Firmicutes was higher in migrants than residents. Firmicutes is a dominant bacterial phylum in the Canada geese gut microbiome in our sample and other studies (Drovetski et al., 2018). However, Firmicutes is a large and diverse phylum, so examining which genera and ASVs were higher in migrants can be more useful for elucidating any possible physiological or functional shifts. Three genera within Firmicutes that were significantly more represented in the migrant group have implications for gut function: Terrisporobacter, Turicibacter, Cellulosilyticum. Bacteria in Terrisporobacter are known to produce acetate (Gerritsen et al., 2014), a compound found to be an important metabolite for skeletal muscle systems in mammals (Frampton et al., 2020). It is possible that bacteria in Terrisporobacter have a role in providing energy for avian skeletal muscle during migration. Turicibacter includes species that metabolize maltose and produce lactate (Bosshard et al., 2002). This may suggest that it aids in breaking down intermediary sugars in the gut, although more research would be needed to elucidate its role in the microbiome. Bacteria in Cellulosilyticum metabolize a variety of carbohydrates, most notably cellulose and cellobiose (Cai and Dong, 2010). As cellulose makes up 25–50% of grass biomass (Sun and Cheng, 2002), and grass is the primary food source for geese in urban environments (Conover, 1991), it is likely that Cellulosilyticum aid goose digestion. Most of the significant taxa at the ASV level were simply subsets of significant genera, although notably, three ASVs of Clostridium sensu stricto (Cluster I) were significantly higher in migrants. Clostridium sensu stricto bacteria are known to metabolize a variety of compounds including carbohydrates, amino acids, and alcohols, and therefore may play an important part in digestive function (Gupta and Gao, 2009; Alou et al., 2017). In total, the taxa that differed significantly between goose populations at the phylum, genus, and ASV levels provide preliminary evidence that migrant geese are characterized by more bacteria that may contribute to the digestion of key dietary components. Whether this pattern results in nutrition and health risks in then populations of urban resident Canada geese or simply indicates adaptive microbial shifts in response to diet remains to be seen. Even in other bird species, these relationships are not well studied. For example, resident curlew sandpipers occupying a less urban environment have a higher abundance of Firmicutes relative to migrants, potentially in response to diet differences (Risely et al., 2017). Nevertheless, diet was not well quantified, and impacts on health were unknown.

Our results highlight differences in microbiome composition and diversity among migratory and resident urban birds. However, as implied above, the precise mechanism that drives microbiome shifts in human-dominated landscapes has yet to be untangled, with diet, stress, and pollution representing the potential competing or synergistic factors (Fuirst et al., 2018; Murray et al., 2020). Further, little is known about how microbes are transmitted among flocking birds, which may lead to non-independence among birds in the same flock. We believe it is unlikely that our results (i.e., that migrants exhibited less inter-individual variation) are due to migrants flocking together because they arrived back in Chicago on different dates to different sites and were not observed together during sampling events. Some of the migrants did use the same sites as residents but if this affected our results, it would have diminished rather than exacerbated any differences in microbiome composition between residents and migrants. Further, the traits of individual birds such as body condition or age may mediate the relationship between migration and urbanization on the microbiome. Future studies that sample social groups in multiple species simultaneously at multiple time points along urban to rural gradients could reveal demographic, social, and landscape drivers of microbial diversity.

The lower microbiome diversity seen in migrants, taken without context, might suggest that the migrant microbiomes are in poor health. However, while reduced microbial diversity is often associated with negative health consequences, this is not always the case. For example, low gut microbiome diversity is associated with a healthy state in human infants (Duar et al., 2020), and higher microbiome diversity has been associated with worse health in urban coyotes (Sugden et al., 2020). Migration can induce a drastic change in the digestive physiology in birds, reducing gut size and requiring birds to metabolize fat and protein reserves (McWilliams and Karasov, 2001). It is possible that the changes in the microbiome may promote energy storage and fat deposition through endotoxins produced by gram-negative bacteria or exotoxins produced by some gram-positive bacteria that provoke an inflammatory response (Risely et al., 2017). Thus, the stresses of migration may shape the microbiome in predictable ways, leading to less inter-individual variation within migrants. The relatively high similarity in the microbiomes of migrants compared to residents suggests a more stable microbial composition and a potential convergence of microbiome composition and function indicative of adaptation. In contrast, the relatively high inter-individual variation in resident geese could signal reduced microbiome resilience, increased microbiome stochasticity, and potentially associated health risks. The “Anna Karenina Principle” (Zaneveld et al., 2017) posits that unstable microbial communities tend to be each unstable in their own unique way, resulting in increased inter-individual variation among “unhealthy” individuals. The patterns we observed in our data are consistent with these dynamics.

Future studies should more directly measure the impact of these microbial shifts on host health. For example, in house sparrows (Passer domesticus), a shift in the microbiome of urbanized populations correlated with lower metabolic gene expression in the microbiome (Teyssier et al., 2018), which could ultimately affect fitness via impacts on processes such as digestion and nutrient intake. Also, urbanization has been linked to a decrease in Firmicutes relative abundance in bird gut microbiomes (Murray et al., 2020), which could impact digestion and nutrition. Given the lower relative abundances of Firmicutes that we observed in resident birds in our study, it seems that the stresses of urbanization may be affecting resident geese more strongly than migrants, who often only winter in urban areas. Quantifying the habitat selection of migrant and resident urban birds along the urban gradient in relation to microbiome composition and host physiology could help identify the consequence of urbanization and migration for wildlife health.

In conclusion, we identified differences in microbiome composition and diversity between migrant and resident Canada geese in urban areas. Determining the relative importance of environment, diet, and physiology in driving these differences, and whether either migrants or residents are healthier, represent important avenues for future research. For example, the differences we detected in the gut microbiomes of migrant and resident Canada geese could have broader implications for the health of migratory bird populations. As urban areas expand around the world, more migratory bird populations may become urban residents (Bonnet-Lebrun et al., 2020; Soriano-Redondo et al., 2020) due to pressures from climate change or the potential benefits of cities such as reliable resources and warmer temperatures. However, the potential benefits to these “colonizing” residents could develop disordered gut microbiomes as a result of altered diets and habitat use. Because microbes affect immune function, including helping hosts resist infections (Harris et al., 2009), a greater understanding of migratory bird microbiomes could be one pathway to improve the chances of success in conservation efforts (Amato, 2013; Murray et al., 2020). Novel techniques that restore natural microbiodiversity could hold meaningful, but relatively untapped conservation potential (Amato, 2013; Teyssier et al., 2018; Wu et al., 2018; West et al., 2019). Beginning to integrate gut microbiome data into more studies of migratory bird ecology represents an important first step toward these novel approaches.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: SRA, PRJNA774835.

Ethical review and approval was not required for the animal study because fecal samples were collected non-invasively from geese that had been GPS-tracked as part of a separate study by different investigators.

MM designed the study and collected biological samples. SO and MS performed laboratory work. SO performed statistical analysis and wrote the original manuscript draft. KA supervised laboratory work and statistical analysis. MM, KA, and MS provided edits. All authors approved of the final version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.