Study area and sampling locations

We conducted our study in the region known as "El Bajío". El Bajío is an urban region (sensu Forman [1]) which, apart from presenting shared eco-physiographic characteristics [48], delimits a Mexican territory with a common socioeconomic history [49]. It is located within the biocultural region of Western Mexico (sensu Schöndube [37]) in the Mexican states of Guanajuato and Michoacán [50]; Longitudes: 101.4077–100.8015° W; Latitudes: 19.63878–20.73814° N; Fig 1). It presents an altitude range from 1721 to 2181 masl [51], with a dry-winter subtropical highland climate (Koppen climate type: Cwb; average annual precipitation: 723.7 ± 92.5 mm; mean annual temperature: 17.8 ± 1.1°C; maximum temperature: 32.4°C; minimum temperature: 2.8°C [52]).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. El Bajío is a biocultural region that presents a continuous human occupation with agricultural activities for the last 3,000 years.

As a result it presents a complex landscape mosaic characterized by high levels of human disturbance. A) The cartographic codex “Mapa de las villas de San Miguel y San Felipe de los Chichimecas y el pueblo de San Francisco Chamacuero” (1578–1580) shows that most of the territory (~70%) was already occupied by crops and cattle production by the end of the XVI century. B) Painting “Paisaje del Bajío” (1960) by the landscape artist Jesús Gallardo™ Carrillo (Republished from [53] under a CC BY license, with permission and courtesy from the owners of “Jesús Gallardo”™, original copyright 1960). This painting depicts the moment in the mid 20th at which most wildland remnants disappeared from the lowlands of the region, intensifying the long-term anthropogenic transformation that El Bajío has undergone. This generated a landscape with large open areas of lowlands used for agricultural purposes and a few remnant areas of native vegetation present on the hills. C) Map of our studied localities in El Bajío (Data by OpenStreetMap under ODbL were used to create the map [54]). We sampled thirteen human settlements: Morelia (1), Irapuato (2), Salamanca (3), Moroleon-Uriangato (4), Zinapécuaro (5), Queréndaro (6), Indaparapeo (7), Francisco Villa (8), Belisario Domínguez (9), San Lucas Pío (10), José María Morelos (11), San Pedro de los Sauces (12) and Colonia Guadalupe (13). The inserted Landsat-8 satellite image shows the location of sampling points for Zinapecuaro to exemplify their general arrangement of sampling sites in the three environments for each location. In addition to each surveyed human settlement (gray), we surveyed the wildlands (green) and productive (orange) environments that surrounded them.

https://doi.org/10.1371/journal.pone.0311290.g001

El Bajío has been an important agricultural region for the last 3,000 years. This process started after the introduction of maize to the region by the Chupicuaro culture ([37]; Fig 1A). Later in the 17th century, it was intensely deforested under Spanish colonial rule to use the wood in the mine´s furnaces, and to open space for cattle grazing areas and cereal production ([39]; Fig 1A). In the mid 20th century, government subsidies and programs incentivized agro-industrial farming that are still present ([36, 39]; Fig 1B). Finally, the North American Free Trade Agreement (NAFTA) promoted the industrialization and urbanization of the region after 1994 [55]. This urban growth has occurred mainly on agricultural areas. Wildlands at our studied region are scattered patches of subtropical xerophytic scrubland, deciduous oak, pine-oak forests, and transition stages among these vegetation types [56]. This environment covers 19.6% of the territory of our study area [57]. Wildlands patches have the most complex vegetation structure in the region, mainly occurring in steep slopes and ravines, where agricultural activities are limited by topography and water availability [55, 57]. The most common native species of trees and shrubs include: Bursera fagaroides, Casimiroa edulis, Celtis pallida, Ehretia latifolia, Fraxinus uhdei, Ipomea murucoides, Opuntia tomentosa, Pinus devoniana, P. leiophylla, P. pseudostrobus, P. laevigata, Quercus castanea, Q. rugosa, Salix humboldtiana, Stenocereus queretaroensis, Vachellia farnesiana, Vackellia pannatula, and Yucca filifera [56]. Human pressure on this environmental condition includes illegal logging; wood gathering for fuel; free-ranging livestock (goats and cows); human-generated fires; the introduction of exotic plant species (Casuarina equisetifolia, Eucalyptus spp., Morus alba, and Schinus molle); and land use change into agricultural and urban conditions [58]. Most wildlands remnants do not have a protection status, with only a few small sized protected areas existing in the region (i.e., Cerro del Punguato, Cerro de Arandas, Siete Luminarias; CONABIO [59]).

The productive environment comprises agricultural activities (crops, greenhouses, plantations, and livestock [55]), covering 73.3% of the territory [57]. Most cultivars are produced using irrigation (open channels with flooding or aspersion systems), with seasonal crops being produced on slopes during the rainy season [56]. Traditional crops include corn, sorghum, wheat, legumes, berries, alfalfa, and vegetables like broccoli, cauliflower, asparagus, and tomatoes [55, 60]. However, in the last decade, several of these cultivars have been replaced with blue agave (Agave tequilana) plantations to produce tequila and mezcal. Crop fields tend to be divided by living fences of both native and exotic tree species (Casuarina equisetifolia, Celtis pallida, Eucalyptus spp., Fraxinus uhdei, Morus alba, Salix humboldtiana, Schinus molle, Vachellia spp.; personal observation R. P-M; [61]).

Finally, the urban environment is constituted by human settlements of different sizes (~0.28 km²—~115 km²) and human densities (~1,067 hab/km²—~9,687 hab/km²; [62]) that cover 6.8% of our study area [57]. The human settlements included in our study were: Colonia Guadalupe, San Pedro de los Sauces, José María Morelos, San Lucas Pío, Belisario Domínguez, Francisco Villa, Indaparapeo, Queréndaro, Zinapécuaro, Moroleon-Uriangato, Salamanca, Irapuato, and Morelia (Fig 1). The majority of them can be classified as residential towns with agricultural pursuits. However, the large urban centers (size ≥ 10 km²; Salamanca, Moroleon-Uriangato, Irapuato, and Morelia) include important commercial and industrial activities [55]. Residential areas display mostly one or two-story houses with little gardens or patios, while several-story buildings are rare and only present in large urban centers. Streets tend to be well illuminated, with exposed electric and communication infrastructure (poles and cables). This often limits the presence of trees and their size when they are planted on the streets [63]. Urban vegetation is scattered inside urban settlements and includes both native (i.e. Celtis pallida, Cupressus lusitanica, Fraxinus uhdei, Salix humboldtiana, and Taxodium mucronatum) and exotic species (Casuarina equisetifolia, Eriobotrya japonica, Eucalyptus spp., Ficus benjamina, Grevillea robusta, Jacaranda mimosifolia, Ligustrum lucidum, Magnolia grandiflora, Schinus molle). It is important to acknowledge that urban vegetated areas present an oversimplified understory relative to wildlands [43].

Bird surveys

We surveyed migratory birds in thirteen localities of El Bajío. Our sampling focused on the southern area of this region. We chose this geographic area because it presents the larger cover of wildlands at El Bajío. Each sampling locality included an urban environment and its adjacent productive, and wildland environments (see our survey example in Fig 1). To determine bird species richness and abundances, we conducted 10 min long point-counts with distance estimations [64]. One observer performed all the bird surveys (A. C.-M.). We recorded birds during a period of 4 hours (from sunrise to mid morning; ~7:00–11:00), targeting the peak activity period of birds. While performing our surveys, we recorded only the birds using the habitat (not those that only flew over the vegetation). All migratory birds detected during the point-counts were included in our species richness analyses. Only migratory birds detected within a 40 m radius of the observer were included in our relative abundance analyses. The radius distance was selected based on the area that included ≥75% of all detections throughout our study (37.5 m radius). We sampled during the winter months after migratory birds had established their overwintering territories in this urban region (November 2018 to February 2019).

To conduct bird surveys, we delimited the areas occupied by each environment on the landscape at each location using a geographical information system [65]. We randomly deployed survey stations inside the delimited area of each environment per location, depending on its size (7–10 survey stations separated by a minimum distance of 250 m). In urban settlements larger than 10 km² we sampled more survey stations (20 in Salamanca, Moroleon-Uriangato, and Irapuato, and 30 in Morelia). In those occasions, survey stations were located at a minimum distance of 500 m. As a result, the total number of sampling stations varied among environments (106 in wildlands, 130 in productive, and 175 in the urban environment). During field work we sampled one locality at a time, sampling a maximum of 10 sampling stations/day. Each sampling station was visited only once. Sampling was conducted on open public spaces or from public roads and streets. When we sampled private spaces we verbally asked for permission to conduct our work. We did not sample when these conditions were not met.

Birds’ functional traits

Birds sampled were classified as migratory following Berlanga [66]. Additionally, bird species that presented both migratory and resident populations (i.e., Troglodytes aedon, Polioptila caerulea) were considered migratory if they were absent during summer or their abundance doubled during the winter in West Mexico (R. P-M. and J. E. S. unpublished data). We excluded migratory waterfowl and raptors from our analyses. Functional traits were classified into categorical or continuous traits. Categorical traits included: 1) their primary trophic guild (carnivore, frugivore, granivore, insectivore, nectarivore, omnivore, and scavenger; [67]), 2) their primary foraging behavior (hawking, excavating, gleaning, and hovering; [67]), and 3) their preferred habitat density (open, semi-open and dense habitats; [68]). Continuous traits included: culmen (mm), beak width (mm), beak depth (mm), tarsus length (mm), tail length (mm), hand-wing index, and body mass (g). We also included the Primary Foraging Stratum mainly used as a continuous variable in our analyses (1: ground, 2: understory, 3: mid-high, 4: canopy, and 5: aerial; [69]). These variables were collected from the AVONET database [68], except the Primary Foraging Stratum. We determined the value of this variable for each species based on the percentages of use for each stratum category in the Elton traits database [55]. We determine the value for each species as the strata at which it accumulated more than 50% of its time, going from the lowest to the highest stratum.

Data processing and analysis

We contrasted the migratory bird assemblages among the three sampled environments of El Bajío region. We compared the assemblages’ species richness, abundance, structure, composition, and functional traits. We calculated migrant bird species richness by interpolating point-count incidence-based rarefactions to 106 points. This number represents the minimum point-count effort for one of our sampled environments (wildlands; [70]). We conducted this analysis using iNext package in R [71]. Estimated species richness values were calculated with asymmetric confidence intervals of 84%. These confidence intervals allowed us to determine non-overlapping comparisons with a p(ɑ) ≤ 0.05 [72]. We contrasted migrant bird abundance among habitat types using a negative-binomial generalized linear model. This family model was chosen over a Poisson one due to the overdispersion found in the latter. We performed pairwise Tukey tests for each model through the "emmeans" R package [73].

We also conducted analyses by trophic guild. We focused on the two dominant trophic guilds of migratory birds at our study region: insectivores and granivores (see below in the results section). Our trophic guild analyses included 1) a comparison of bird species richness, relative abundance and functional diversity for each trophic guild (insectivore or granivore) among environments; and 2) a comparison of abundances among the different groups that constitute the selected categorical traits (tropic guilds, primary foraging behavior and preferred habitat density) in the three environments. These comparisons were conducted similarly to those mentioned above for all migratory birds.

We evaluated the structure and composition of migratory bird assemblages by calculating the effort-based relative abundance of each species in each environment. We standardized all species’ relative abundances to individuals per point-count to achieve this (individuals/point-count). We used these abundance values to generate rank-abundance plots following Whittaker [74]. Briefly, a rank-abundance plot displays the assemblage evenness by ordering the species on the horizontal axis by their relative abundances from highest to lowest and displaying their relative abundances on the vertical axis. We used rank-abundance plots to compare the assemblage’s evenness among environments. We compared them using a linear model with the log10 transformed relative abundances as the response variable and the environment and their species ranks as interacting factors.

We assessed the migrant bird assemblage composition by calculating a Bray-Curtis dissimilarity index considering all locations [75]. We subsequently processed them by using a non-metric multidimensional scaling (MDS) on "vegan" R package [76]. Then we compared the assemblages among environments with an adonis test [76]. We considered the environment as the only factor, and we performed a subsequent pairwise posthoc test [76]. We also assessed each bird species association with the three environments by calculating phi coefficients with all recorded species [77]. The phi coefficient ranges from -1 to 1, where 0 represents no species association with the environment type. A positive coefficient reflects a positive association, and a negative coefficient reflects a negative one [77]. The obtained phi coefficient is independent of sample size and was calculated using "indicspecies" package in R [78].

We assessed differences in functional diversity among environments using multitrait Gower distance and community-weighted mean values (CWM) considering all functional traits (see above). These variables were weighted by the species’ relative abundances in each environment [79]. Given that the Gower distance values can give disproportional weight to certain variables, we calculated them and their "gawdis" R function following de Bello et al. [80]. Afterward, we used the calculated Gower distance to obtain the FEve, FDiv, and FDis functional diversity indexes [79, 81]. We calculated CWM with “FD” R package [82]. The functional diversity index values obtained from functional traits comprised of continuous values were compared using generalized linear models considering the environment as the only factor. Categorical functional traits were also contrasted with generalized linear models considering the environment and the categorical trait as two interacting factors. In both cases, we assessed the differences given by the model using a posthoc pairwise Tukey test by their estimated marginal means [73]. We assumed relevant contrasts at a p(ɑ) = 0.05.

Changes in migratory bird assemblages associated with land use change in El Bajío

As a result of 3,000 years of human management, our studied region presents one of the lowest levels of ecosystem integrity in the country [83]. Associated to the long history of intense human activities in this region, most of the migratory bird species we recorded do not present conservation problems [66]. Only one species, the long-distance migratory hummingbird Selasphorus rufus is considered as near threatened (IUCN red list). Interestingly, this species was only present on wildlands, the environment that presented the highest number of migratory species in the region.

The wildland patches of El Bajío contain a significant number of migratory bird species existing in other areas of Western Mexico with similar environmental characteristics (same altitudinal and latitudinal range; see S1 Fig; [84]). In this environment, we recorded a number of migratory bird species that did not differ from the calculated species richness for the regional pool. Most of the species that we only found in this environment prefer dense and semi-open habitats, and are associated with forests and shrublands as their primary habitat type. Additionally, this group of migratory birds exclusive to wildlands may suggest a shared low tolerance for human activity. However, migratory bird species unique to the wildlands of El Bajío, such as Turdus migratorius and Geothlypis tolmiei overwinter in neotropical cities of other regions of Mexico (S1 Fig and S1 Dataset; [42, 44, 84]), which suggest that the same species can respond differently to human activities depending on the regional context. These results indicate that despite its long historical alteration by humans, the wildlands of El Bajío are an essential habitat for overwintering non-endangered migratory birds today [32, 85, 86].

El Bajío shows two different trajectories of land use change that have affected migratory bird assemblages. The first one is the transformation of wildlands into productive areas. This transformation started millennia ago with the arrival of agriculture [87], and has undergone an important increment in recent centuries through the opening of extensive areas for intensive cultivation (Fig 1A and 1B; [49]). This first environmental change due to human activities generated essential changes in the assemblages of migratory birds, leading to a 40% species loss (13 species), with most of the missing species being those associated with semi-open and dense habitats [68]. Yet, productive environments offered new habitat conditions that allowed the arrival of eight migratory species with open habitat preferences to this region (21.8% of the total species richness observed in wild environments). These species could have invaded El Bajío from the grasslands of the Central Altiplano located north of it (≈100 km from our study area). The combination of species losses and turnover generated an assemblage that is 24% (Bray-Curtis: 0.22–0.26; CI 95%) similar to the one present before this agricultural land use change occurred.

The migratory bird assemblage of the productive environment also differed in its abundance in relation to wildlands, presenting 40% fewer individuals. This reduction in abundance was not associated with the loss of some species that accompanied the conversion of wildlands to productive environments. These species were already present in low abundances in wildlands, representing only 8.2% of the total migratory bird abundance in that environment. However, it was caused by a reduction in the number of individuals of two insectivorous species that were abundant in wildlands: Setophaga coronata and Polioptila caerulea that together represent 51.7 and 22.9% of the total abundance in wildlands and productive environments, respectively (Fig 3). The abundance difference of these species in these environments suggests that crucial resources may limit their densities in productive areas [88]. Interestingly, while some open-habitat species were dominant in the productive environment, their numbers did not compensate for the abundance loss compared to wildlands. While the productive environment of El Bajío offers opportunities for the arrival of open habitat migrant birds, it harbors assemblages that present lower species richness and abundance values, limiting their conservation potential value in the region [27].

The second trajectory of land use change in El Bajío is associated with urbanization transforming productive environments [49, 88]. This land use transition modifies the assemblage of migratory birds by losing 42.3% of the species that use the open areas (11 species). These species represent 39.5% of the migratory bird abundance in the productive environments of our study region, indicating that urban areas offer different habitat elements and food resources to this group of birds [47, 89]. For example, migratory birds in urban environments are strongly associated with green infrastructure like large street trees, or parks where trees are common habitat elements [36, 38, 42]. The reduction in bird abundance caused by the urbanization of productive environments was mainly associated with the loss of two species of migratory birds from open areas that were exclusive to, and abundant, in productive environments. These two species were Tachycineta bicolor and Passerculus sandwichensis, which together comprised 30.7% of the total abundance in productive environments. These results also suggest that the lower abundance of migratory birds in urban areas may be caused by reduced areas of high quality urban habitat and limited food resources in this environment [90].

Urbanization processes have also occurred in areas that used to be covered by native vegetation [89]. While this land use change occurs less frequently, in the last half century this process has generated housing suburbs of different socioeconomic levels around the largest cities of the region (e.g., Morelia, Irapuato, Moroleon-Uriangato, and Salamanca; [91]). This transformation of wildlands due to urbanization represents a loss of 43% of the species of the migratory bird assembly. Their populations also declined, as wildlands had 250% more individuals. These declines in species richness and abundance resembles those previously described by MacGregor-Fors [36] for the city of Morelia.

Interestingly, the urban migratory bird assemblage shared three of the 13 species present in wildlands that were absent in productive environments. This indicates that some migratory birds associated with wildlands are able to use urban environments [38]. Zuckerberg [92] observed that during their stage of long distance movements, migratory birds present a higher plasticity in habitat choice than during their breeding period, with some species favoring urban environments. Furthermore, during the winter, urban areas can be used by some migratory birds considered as interior forest specialists [38]. Our results, together with those from other studies, indicate that urban green features, like large trees, can offer migratory birds habitat elements that function similarly to some of those present in wildlands.

Differences in assemblage structure among environments

The main differences in the migratory bird assemblages of the three environments were associated with: 1) changes in the abundances of common species shared among the three environments; and 2) the presence of some species that take advantage of particular habitat elements and resources offered by productive and urban environments. Similar mechanisms have been suggested to explain the spatial distribution of migratory birds in other Neotropical human-transformed regions [22, 24, 35]. For example, the abundance of specific food items, like grain or small arthropods, have been associated to higher densities of granivorous and insectivorous migratory birds respectively [27, 30, 32]. Additionally, birds from these two trophic groups benefit differentially from complex vegetation, with granivores depending on areas that present a dense understory with a high diversity of herbaceous plants, and that present low numbers of trees [27, 30, 92]. While insectivorous birds increase their species richness and abundances in response to the complexity of the canopy [22, 36, 42, 44].

While migratory bird assemblages differed among environments, they shared an important number of species. This shared group included 15 species, and represented 34% of the regional pool of migratory bird species at El Bajío. Shared species constituted 47% of the wildlands bird assemblage, while these values were larger in productive and urban environments (57% and 68% respectively). Most of the shared species were small insectivorous birds, mainly warblers (Fig 3 and S1 Dataset). These species have been stated to have a broad capacity to take advantage of a large diversity of habitats [26–28]. They have also been mentioned as being common or abundant in Neotropical urban bird assemblages [42].

The core group of migratory bird species shared among the three environments includes dominant species that represent 82% of the total regional abundance (Fig 3). While the shared species included the most abundant species, it is essential to notice that they differed in their abundance among environments, defining the structure of the different bird assemblages. Species abundance differences among environments have been associated with a change in the abundance of food resources and the quality of habitat patches in productive and urban environments [38]. It has also been associated with a gradient of human activity [93]. Assessing the role that habitat elements and human activity have on migratory bird behavior and habitat use offers important insights to manage this group of birds in human modified environments. In consequence, we recommend future studies to evaluate the role that trees (species involved, their architecture, and food resources), crops (type and managing activities), and human activity (human presence, noise, mobility) play on the behavior and ecology of migratory birds in both productive and urban overwintering environments [14].

Functional traits of migratory birds in El Bajío

The regional migratory bird assemblage was primarily composed of insectivorous species, followed in importance by granivorous birds. This result is consistent with findings from other Neotropical overwintering regions [27, 28, 32]. Migratory bird assemblages varied in specific functional traits among the region’s studied environments [12, 93]. However, the core of species shared among environments (see above) suggests the existence of critical common functional traits. As a result in today’s landscape, any of the region’s environments can maintain the most abundant functional traits of the regional pool of migratory birds.

The core of shared species was characterized by small body-size insectivorous birds that engage in gleaning foraging behavior, and show preferences for dense and semi-open habitats. These species have been reported as typical winter inhabitants of Neotropical wildlands [27, 28, 32, 92], urban areas [38, 42], and productive environments [27]. Their presence and abundance in anthropogenic environments have been associated with a complex vegetation structure that includes a high density of trees, and an intricate lower stratum of shrubs and herbs [27, 38, 43]. While this core of species has the potential to maintain stable populations in anthropogenic environments [43, 85], it is crucial to understand their interactions with specific habitat elements, such as specific trees and shrub species, to perpetuate their populations in the future.

Other species in this core group included sparrows, cardinals, an icterid, and a hummingbird (see S1 Fig and S1 Dataset). This diverse collection of trophic groups of migratory birds in productive and urban environments highlights the broad spectrum of resources available to them in human-altered sites [94]. This high trophic diversity could be related to the presence of flowers and fruits in urban sites, and seeds in productive environments [27, 30, 31, 35, 92]. Additionally, human-modified habitats present novel habitat features, like buildings, cables, poles, and other structures, that help support bird assemblages associated with structurally complex habitats [39, 41, 95]. The presence of a surplus of water in human-altered environments during the cold-dry season of the year, when water is limiting, could also promote the presence of a diverse array of trophic guilds in these environments [17, 44, 95, 96].

The differences in migratory bird assemblages across environments, based on their functional traits, allow us to identify crucial environmental features that can act as opportunities or limitations for migratory birds. Wildlands presented abundant insectivorous birds, with our findings confirming the prevalence of this guild in Neotropical wildland assemblages [27, 28, 32]. However, the ten migratory bird species we only recorded on wildlands were diverse in their diets (nectarivores, insectivores, and omnivores) and presented semi-open and dense habitat preferences (S1 Fig and S1 Dataset). Their presence indicates an affinity for a complex vegetation structure that includes mature trees and several layers of vegetation (herbs and shrubs) while simultaneously offering semi-open spaces combined with dense understory areas [32, 38].

The migratory bird assemblage of productive environments was functionally characterized by a balanced proportion of the abundance of insectivorous and granivorous birds, and balanced abundances by habitat preferences (dense, semi-open, and open habitats). However, the seven migratory birds exclusively recorded in productive environments had open and semi-open habitat preferences. Their functional traits underscore the importance of open spaces for aerial foraging in productive environments (e.g., Tachycineta bicolor) while also promoting birds that forage for seeds on the ground (e.g., Passerculus sandwichensis; [27, 30, 31].

Our results indicate that the productive environment harbored the regional pool of migratory granivorous birds. Similarly, granivorous migratory birds have been described as abundant in intensively managed crops during their overwintering period in other regions of the continent [27, 30, 92]. Associated with this, the migratory bird assemblage of productive environments had the highest beak depth and the lower primary foraging stratum of our studied region. These functional traits indicate the importance of crop-related food resources and habitat characteristics of productive environments for some migratory bird species [27, 31, 35].

We also found that insectivorous birds had similar abundances between productive and urban environments. Both environments possess scattered tall trees that benefit this trophic group. However there are important differences among the way trees are distributed in space in these two environments [97]. In the productive environment, trees are primarily utilized as living fences marking property limits and being surrounded by crops. In urban environments trees are mostly concentrated in parks, and scattered mainly along streets. Our results indicate the importance of living fences and urban trees in Western Mexico’s, as they increase the vegetation complexity of intensely human modified environments, offering opportunities for birds that do not feed on seeds or use the ground for foraging [27, 30, 31, 97].

We were surprised to find a lack of migratory granivorous species in the urban assemblage. This is remarkable if we consider that resident urban bird assemblages are typically dominated by granivorous species, suggesting the presence of abundant food resources for this trophic guild [8, 98–100]. We propose three non-mutually exclusive hypothesis to explain the absence of granivorous migratory birds in the urban environment of El Bajío: 1) competitive exclusion by abundant non-migratory granivorous birds, which includes exotic species like the House sparrow (Passer domesticus), a species that is aggressive towards smaller granivorous bird species [8, 100]; 2) the lack of understory vegetation in urban environments that could limit grassland migratory granivorous species that forage and hide from predators among dense herbs [93, 97]; and 3) a lack of behavioral and physiological traits in migratory granivorous birds that prevent them from tolerating the urban stressors associated to human activity [39, 41, 101].

The lack of granivorous birds in the urban environment generates an assemblage dominated by insectivorous birds. Interestingly, the migratory insectivorous species using urban areas were characterized by presenting dense habitat preferences, higher Primary Foraging Stratum values, and smaller beak dimensions than those species using the other environments (S1 Text; Fig 6). As hypothesized by Greenberg’s breeding currency hypothesis [28], the availability of small arthropods in the tree canopy and plant foliage could be enhanced in environments that presented depauperized bird assemblages, like urban areas [32, 43, 85]. This could promote an abundance of insectivorous migratory birds with short culmens like the one we found. The three migratory birds exclusively recorded in urban environments had dense and semi-open habitat preferences and were associated to water related habitats (Parkesia noveboracensis, Setophaga petechia and Vireo plumbeus; S4 Dataset). This opportunity could arise by the urban year-round water surplus [15, 44, 95]. The functional characteristics of the urban migratory bird assemblage of El Bajío suggests that urban areas can act as simplified forests with a high water surplus [43]. Thus, our results indicate that offering water sources and increasing tree abundance and vegetation complexity is essential for enhancing migratory bird populations in urban environments.

Future opportunities for migratory birds in a Mexican urban region

The high spatial dispersal ability of migratory birds makes them an ideal model for understanding how different factors mold their assemblages in the human-modified landscapes. Our results indicate species composition differences among environments, supporting the hypothesis that migratory species are filtered by their functional traits in anthropic environments [19, 20, 31, 93]. Additionally, our results suggest that while some migratory species can use the three studied environments, their lower abundances in the human-modified ones are related to a limitation on the abundance of key resources [22, 32, 102–104]. The shared core of migratory bird species indicates that some habitat features, primarily tree density and size, positively contribute to their presence and abundance in any environment. This result offers a promising opportunity to manage this biological group in a highly modified region.

Sadly, human modification of the wildlands of El Bajío by agricultural activities like grain or legume production is not the worst scenario migratory birds face in West Mexico today. In the last decade, traditional crops (corn, wheat, oats, sorghum, alfalfa, and beans; [56]) have been replaced by blue agave cultivars (Agave tequilana) used for the production of tequila [105]; personal observations). Agave production is not only being conducted on flat areas that present irrigation, but also on the hillsides where the last remnants of wildlands are located [105]. Agave cultivars have a low vegetation complexity, lacking the presence of trees, shrubs, and herbs. As a result, this habitat only supports nine migratory bird species (Ceja-Madrigal and Schondube unpublished data). This change from previous crop systems could generate additional negative impacts on the migratory bird populations of this region. As a result, we recommend conducting studies on the effects of the dramatic expansion of blue agave production in Mexico on migratory bird populations.

While conservation efforts for migratory birds in this region should focus on maintaining wildland patches [106], it is crucial to understand that both productive and urban environments offer opportunities to promote their conservation inside highly human modified regions [107]. The accelerated rate of land use change to agave cultivars experienced by this region suggests that urban areas can act as an essential habitat for migratory bird species as this productive system expands over the landscape [105]. To achieve this, we need to manage cities so they can offer good quality habitat for migratory birds. The fact that several migratory birds can use urban areas for overwintering while maintaining good physical condition indicates the need to integrate these birds into Neotropical urban planning [14, 15, 43, 107]. Our results indicate that increasing habitat quality for migratory birds in cities could become crucial for their conservation, considering the transformation suffered by wildlands and productive environments in the Anthropocene.