Introduction
The study of species diversity has been a main research topic in community ecology (Gaston Reference Gaston2011). In comparison, other ways of characterizing diversity, like functional and phylogenetic, have received less attention, despite being useful in revealing information related to habitat gradients (McGill et al. Reference McGill, Enquist, Weiher and Westoby2006), evolutionary processes and community assembly (Boyce et al. Reference Boyce, Shakya, Sheldon, Moyle and Martin2019; Montaño-Centellas et al. Reference Montaño-Centellas, McCain and Loiselle2020), ecosystemic services (Michel et al. Reference Michel, Whelan and Verutes2020) and ultimately biological conservation (Winter et al. Reference Winter, Devictor and Schweiger2013; Tucker et al. Reference Tucker, Aze, Cadotte, Cantalapiedra, Chisholm, Díaz, Grenyer, Huang, Mazel and Pearse2019).
The Neotropical Seasonal Dry Forests (hereafter NSDFs), which range from western Mexico to northern Argentina, are among the most biodiverse vegetation types in the American tropics. NSDFs originally covered roughly 40% of the region, but now are underrepresented in the Natural Protected Areas systems and suffer high deforestation rates (Prieto-Torres et al. Reference Prieto-Torres, Rojas-Soto, Bonaccorso, Santiago-Alarcon and Navarro-Sigüenza2019). The physiognomy of the NSDFs and their prolonged dry season, when most tree species lose their leaves, represents a challenge for the species inhabiting them. At the same time, such seasonal changes allow the existence of diverse ecological niches, favouring a high alpha and beta bird species diversity in Southwestern Mexico (Rowley Reference Rowley1966; Hutto Reference Hutto1985; Binford Reference Binford1989; Hutto Reference Hutto1994; Vázquez-Reyes et al. Reference Vázquez-Reyes, Arizmendi, Godínez-Álvarez and Navarro-Sigüenza2017), Central and South America (Gillespie and Walter Reference Gillespie and Walter2001; Gonçalves et al. Reference Gonçalves, Cerqueira, Brasil and Santos2017; Avendaño et al. Reference Avendaño, López-O and Laverde-R2018; Vásquez-Arévalo et al. Reference Vásquez-Arévalo, Grández-Casado, Muñoz-Pizango, García-Villacorta and Gagliardi-Urrutia2018).
According to Becerra (Reference Becerra2005), the NSDFs arose between 30 and 20 mya in southwestern Mexico, when both the Sierra Madre Occidental (western Mexico) and the Transmexican Volcanic Belt (Central Mexico) appeared. These biogeographic complex patterns have allowed the prevalence of the dry and warm conditions that currently characterize the region. Furthermore, the climatic shifts during the Pleistocene, characterized by both glacial and interglacial periods, promoted the existence of refugia (Castillo-Chora et al. Reference Castillo-Chora, Sánchez-González, Mastretta-Yanes, Prieto-Torres and Navarro-Sigüenza2021) or isolated extensions of NSDFs where successive episodes of speciation took place (Hubbard Reference Hubbard1973; Rahbek et al. Reference Rahbek, Borregaard, Colwell, Dalsgaard, Holt, Morueta-Holme, Nogues-Bravo, Whittaker and Fjeldså2019). The climatic stability during the late Quaternary shaped the high biological diversity that is currently observed in the NSDFs (Arango et al. Reference Arango, Villalobos, Prieto-Torres and Guevara2021).
Avian diversity associated with NSDFs has been partially addressed using community-level ecological analyses. Yet, this accumulated knowledge is less than that of other tropical vegetation types (Martínez-Ruiz and Rueda-Hernández Reference Martínez-Ruiz and Rueda-Hernández2024). These studies are either descriptive (Rowley Reference Rowley1966; Vázquez-Reyez et al. Reference Vázquez-Reyez, Moya and del Coro Arizmendi2009) or have focused on the differences between primary forest and other anthropized habitats (MacGregor-Fors and Schondube Reference MacGregor-Fors and Schondube2011; Vázquez-Reyes et al. Reference Vázquez-Reyes, Arizmendi, Godínez-Álvarez and Navarro-Sigüenza2017; Alvarez-Alvarez et al. Reference Alvarez-Alvarez, Corcuera and Almazán-Núñez2018). In contrast, the number of studies dealing with the structure and function of ecological assemblages other than taxonomic diversity in the American tropics is limited to a few works mainly performed in South America (Dehling et al. Reference Dehling, Fritz, Töpfer, Päckert, Estler, Böhning-Gaese and Schleuning2014; Araneda et al. Reference Araneda, Sielfeld, Bonacic and Ibarra2018; Martínez-Ruiz and Rueda-Hernández Reference Martínez-Ruiz and Rueda-Hernández2024). Vázquez-Reyes et al. (Reference Vázquez-Reyes, Paz-Hernández, Godínez-Álvarez, del Coro Arizmendi and Navarro-Sigüenza2022) analysed the functional differences between the primary dry forests and human-made habitats, finding that birds in the anthropized environments showed traits of species that follow a ruderal life strategy. At more southern latitudes, overdispersed avian communities are usually reported at low-elevation localities, contrasting with clustering patterns of highland bird assemblages (Dehling et al. Reference Dehling, Fritz, Töpfer, Päckert, Estler, Böhning-Gaese and Schleuning2014). In environmental gradients of the dry forests in Brazil, differential responses of both functional and phylogenetic diversities have been documented, mainly caused by climatic differences and habitat heterogeneity (do Nascimento et al. Reference do Nascimento, Correia, Ruiz-Esparza and Gouveia2018). Due to both the prevailing tropical forests’ landscape configuration and lack of long-term ecological data, early community-level studies have implemented a space-for-time substitution approach.
The dry slopes of Central Oaxaca belong to the Mexican Transition Zone, a biogeographic region characterized by the presence of biotic elements from both the Nearctic and the Neotropical zones and restricted-range species (Morrone Reference Morrone2019). Accordingly, the avifauna of such slopes is composed of a mixture of species of both boreal and austral affinities, as well as endemic species (Peterson et al. Reference Peterson, Escalona-Segura, Zyskowski, Kluza and Hernández-Baños2003). The forests of Southwestern Mexico are important for the number of migratory birds that come from the Nearctic zone (Hutto Reference Hutto1994), influencing the community dynamics in a seasonal manner by increasing the taxonomic diversity (Somveille et al. Reference Somveille, Manica, Butchart and Rodrigues2013).
Despite a complex history of human intervention, the NSDFs have persisted (Castillo et al. Reference Castillo, Magaña, Pujadas, Martínez and Godínez2005). However, they have recently suffered from intensive deforestation to be converted into agricultural lands, cattle pastures or urban areas (Prieto-Torres et al. Reference Prieto-Torres, Rojas-Soto, Santiago-Alarcon, Bonaccorso and Navarro-SigüEnza2019). It has resulted in Mexico’s high rate of deforestation and rapid land use/land cover change (Ceballos and García Reference Ceballos and García1995). Whilst decreeing natural areas has been advised as a main strategy for the conservation of these biodiversity hotspots (Prieto-Torres et al. Reference Prieto-Torres, Nori, Rojas-Soto and Navarro-Sigüenza2021), it may not be the best choice in places with a long history of land use and management as expected in the southern Mexican highlands, inhabited by peasant and indigenous communities (Van Vleet et al. Reference Van Vleet, Bray and Durán2016). Accordingly, alternative conservation schemes may be considered to ensure the coexistence of humans and biodiversity. In addition, the priorities for conservation are often defined considering the species richness (Montaño-Centellas et al. Reference Montaño-Centellas, Loiselle and McCain2021), but other diversity dimensions are also important in places such as the tropical mountains, characterized by a high evolutionary potential (i.e. speciation probability, Voskamp et al. Reference Voskamp, Baker, Stephens, Valdes and Willis2017) and ecosystemic services (Michel et al. Reference Michel, Whelan and Verutes2020).
Herein, we compared the bird diversity of two contrasting vegetation types (Tropical Dry Forest (TDF) and Oak Woodland (OW)) in a tropical highland region. Our specific questions are: (1) How different are bird assemblages at two contrasting vegetational types in terms of structure and function? (2) Are there seasonal patterns shaping bird diversity in this high-elevation region? (3) Is there any association between bird diversity and vegetation type? As follows, we tested a series of well-established community-level ecological hypotheses. First, TDFs would have higher multifaceted alpha diversity than the more humid conditions of OWs since species diversity usually declines with increasing elevation. Second, higher diversity in the non-breeding season than in the breeding season due to the arrival of the Nearctic-Neotropical migrant species (Hutto Reference Hutto1994). Third, we would expect to find a high beta diversity because of the narrow distribution ranges and low dispersal abilities of the tropical species (Montaño-Centellas et al. Reference Montaño-Centellas, Loiselle and McCain2021). Finally, we would expect an ecological association between bird species and vegetation type. We further discuss our findings and the possible explanations according to the evolutionary and ecological information available for those two types of vegetation and their conservation implications. This study delivers much-needed empirical information on birds of the NSDFs.
Methods
Study site—We made the fieldwork in the municipality of San Pedro Totolapam (hereafter Totolapam), Central Oaxaca, Southern Mexico (Figure 1a). Totolapam lies within two biotic regions (Morrone Reference Morrone2019): (1) Sierra Madre del Sur, a mountainous complex that ranges from Jalisco to Oaxaca (Figure 1a); (2) Pacific Lowlands, a broad region that spans from Sonora to Chiapas (Figure 1a). The Sierra Madre del Sur reaches its maximum altitude of 3,720 m a.s.l. in this region. The wet season goes from June to October (CONAGUA-SMN 2024), with a mean annual temperature and precipitation of 24°C and 468 mm, respectively. The main vegetational types were TDF at the lower altitudes and OW at higher altitudes (Binford Reference Binford1989). Other land covers in the area were agricultural fields (mainly milpa and agave plantations), pastures, mines and small urban settlements (Figure 1a). Despite human activities, the forests of the region are in good conservation shape (Figure 1b).

Figure 1. Map of the study site. (a) Land covers and sites in San Pedro Totolapam. (b) Panoramic view of the Neotropical seasonal dry forests in San Pedro Totolapam.
Bird surveys—We visited the study site between January and November 2021. We surveyed birds in five sites distributed along the two vegetation types. The sites in TDF were 2 km away from road 190 that runs from Oaxaca City to Tehuantepec, while the remaining three OW sites were 4 km North of the same road (Figure 1b). TDF sites were located at a mean altitude of 835 m a.s.l. and the OW sites were at a mean altitude of 1548 m a.s.l. We established one transect at each site, deploying 10 points separated by a minimum distance of 250 m to assure statistical independence. We performed 25-m radius point counts along a 2.25 km transect at each site, then we sampled 33.75 ha in OW and 22.50 ha in TDF. We simultaneously carried out two field techniques: (1) point counts and (2) soundscape recordings (Suárez-García et al. Reference Suárez-García, González-García and Celis-Murillo2017). One field technician (OSG) registered all individuals seen or heard within 5 minutes, while recording soundscapes at the same time by using a Telinga cardioid microphone mounted over a tripod, one digital recorder Tascam DR-40 and headphones. Audio files were recorded in WAV format at a sample rate of 24 bits and stored in a hard drive. Both field techniques allowed us to gather species and abundance data. We visited each site at least once every two months in a lapse of 11 months. We accomplished all field surveys under good weather conditions (no rain, no fog, no strong winds) from dawn to four hours onwards.
We established bird migratory status according to the published species accounts (Howell and Webb Reference Howell and Webb1995). We classified restricted-range species as endemics (exclusively distributed within the Mexican territory), quasi-endemics (range including Mexico and less than 35 000 km2 outside the Mexican boundaries) and semi-endemics (range including entirely the Mexican territory in any of their life stages) (González-García and Gómez de Silva Reference González-García, Gómez de Silva, Gómez de Silva and de Ita2003). We defined the biogeographical affinity of birds based on the location of the breeding range of the recorded species (Palomera-García et al. Reference Palomera-García, Santana and Amparán-Salido1994), with breeding range information reported by Howell and Webb (Reference Howell and Webb1995). Such criterion indicated to us the existence of six distributional groups: (1) Nearctic, (2) Neotropical, (3) Endemic, (4) Quasiendemic, (5) Semiendemic and (6) Widespread. Finally, we assessed the species risk status according to the Mexican law (SEMARNAT 2010) and the IUCN Red List.
Data analysis—By using the chisq.test function of the rstatix package, we made chi-squared tests to determine if there were statistical differences in richness at both the family and order levels, making a pairwise post hoc test with Bonferroni corrections. To extract information on bird species and abundances from the audio files, one technician (OSG) analysed the archives in a PC by using Audacity free software. He compared the detected vocalizations with the records available at the Xeno-canto sound library when we had doubts about the species identity. With the data gathered by both field methods, we constructed a species x site abundance matrix. We excluded from the analysis both diurnal and nocturnal raptors (except Laughing Falcon and Ferruginous Pigmy-Owl), swallows and waterbirds because our field methods were not suitable to detect them and reasonably estimate their relative abundances (Suárez-García et al. Reference Suárez-García, González-García and Celis-Murillo2017).
To analyse the diversity patterns, we primarily focused on the 11-month pooled data obtained for both vegetation types. To explore the seasonal diversity patterns, we subdivided our data into three seasons (based on the presence or absence of the migratory species). First season included data from January to April (first migratory period, hereafter “M1”); the second season from May to August (breeding period of all-year residents, no observed migratory birds, hereafter “B”) and the third season from September to November (migratory period 2, hereafter “M2”).
Sample completeness—We calculated the sample coverage (SC) to quantify sample completeness (Chao and Jost Reference Chao and Jost2012). SC is the probability that a new individual found in a studied community belongs to the species already represented in the sample. We calculated (1) the total SC and (2) SC by seasons for each vegetation type. We also used the SC to compare the performance of the two field methods used in this work.
Alpha diversity—We expressed diversity by Hill numbers, according to the approach proposed by Chao et al. (Reference Chao, Chiu and Jost2014). In the case of species diversity, the measured attribute was the effective number of species, in the case of functional diversity was the effective sum of species pairwise distances, and in the case of phylogenetic diversity was the effective branch-length in a phylogenetic tree (Chao et al. Reference Chao, Henderson, Chiu, Moyes, Hu, Dornelas and Magurran2021). In all cases, we considered three different diversity orders: q = 0, no species abundances considered, q = 1, all observed abundances considered, q = 2, only the abundances of dominant species considered. We compared the vegetation patterns of the three diversity types by calculating the asymptotic q diversity profiles with the ObsAsy3D function of the INEXT.3D R package (Chao and Hu Reference Chao and Hu2024).
In the case of functional diversity, we first constructed a trait matrix with data of body characteristics (average mass, total length, sexual dimorphism), feeding strategy, food, seasonality, clutch size and social behaviour (Table S1) following the Birds of the World database (Billerman et al. Reference Billerman, Keeney, Kirwan, Medrano and Sly2025). We then calculated a similarity matrix by using the Gower’s index, which is appropriate to deal with both categorical and continuous data. Finally, by using both the similarity and the abundance matrices, we calculated the functional diversity q profiles. We also constructed a dendrogram of functional groups with the gawdis function of the gawdis R package. For that, we used the gap statistic analysis to define the optimal number of functional groups in our cluster. We compared the abundances between functional groups by making a Fisher’s exact test and pairwise comparisons with Bonferroni correction.
Regarding phylogenetic diversity, we downloaded a phylogeny subset (all our recorded species) of 500 samples from birdtree.org (Jetz et al. Reference Jetz, Thomas, Joy, Hartmann and Mooers2012). Then, we constructed a phylotree object with the maxCladeCred function of the phanghorn package (Schliep Reference Schliep2011). Next, we calculated the diversity values with both the maximum clade credibility tree and the abundance matrix.
For all diversity metrics, we calculated the inequality factor (IF0,2) dividing 0 D by 2 D, with the aim of comparing the dominance between communities. The maximum value of IF0,2 is 0 D (either species, functional or phylogenetic), and the lowest value is 1; in this case, high IF0,2 values indicate high dominance (Jost Reference Jost2010).
Beta diversity and species composition—We analysed the differentiation between avian communities by calculating and drawing beta diversity profiles from order 0 to order 2. Beta diversity was expressed as the number of effective communities. In the case of species and phylogenetic diversity, the minimum possible beta diversity value was one (in the hypothetical case that the two communities were identical in both composition and abundance) and a maximum of two (hypothetical case when the communities were completely different in both composition and abundance), while in the case of functional diversity the minimum possible value was one and the highest was four. We did the analysis with the hill_taxa_parti, hill_func_parti and hill_phylo_parti functions of the HillR package. In addition, we made rank-abundance plots to depict the species composition and community structure.
Species distribution—We performed a correspondence analysis (CA) to assess the distribution of the species across the two vegetation types and seasons. We reported the chi-squared value of the relationships between the species and sites/seasons, as well as the eigenvalues and proportion of represented variance per axis. We excluded only from this single analysis the species represented by less than three individuals in any of the vegetation types or seasons considered, because these observations could have been done at random (Gotelli and Ellison Reference Gotelli and Ellison2004). We assessed the distribution of birds based on their biogeographic affinity in our two studied vegetation types by performing a pairwise Pearson chi-squared test with the pairwise_chisq_test_against_p function of the rstatix R library (Kassambara Reference Kassambara2019) and depicting the relationships by a mosaic plot with the mosaicplot function of the graphics R library.
Results
We recorded 73 species belonging to 22 families and 11 orders (Table 1). The global chi-squared tests at both order (χ2 = 337.76, DF = 10, p < 0.001) and family (χ2 = 56.28, DF = 21, p < 0.001) levels were significant. According to the chi-squared test pairwise comparisons with Bonferroni corrections, Tyrannidae had statistically higher richness than the remaining families (χ2 = 36, DF = 1, p < 0.001), while Passeriformes had a higher richness than the remaining orders (χ2 = 304, DF = 1, p < 0.001). Fifty-seven species were all-year residents, 14 non-breeding migrants and two breeding migrants; chi-squared test was statistically significant (χ2 = 68.7, DF = 1, p < 0.001) and richness of seasonal categories was statistically different to each other (Table S2). We recorded 14 endemics, three semi-endemics and three quasi-endemics; endemics richness was statistically higher than both quasi-endemics and semi-endemics (χ2 = 312.1, DF = 1, p < 0.01). Regarding conservation status, four species are threatened according to the Mexican law and one species (Sumichrast Sparrow Peucaea sumichrasti) is included in the IUCN Red List as Near Threatened.
Table 1. Bird species registered in this study. AB—Abundances, CS—Conservation status, S—Seasonality, E—Endemism, RL—UICN Red List, NI—Not included; Pr—Special protection; A—Endangered; P—Risk of extinction; R—All-year resident; BM—Breeding migrant; NBM—Non-breeding migrant, I—Introduced; E—Endemic; Q—Quasiendemic; S—Semiendemic; NT—Near threatened; VU—Vulnerable

Sample completeness—We made 20 visits to the sites considered in this study (Table S3), completing 1100 minutes of observation and the same amount of sound recordings. We registered 796 individuals of 62 species with sound recordings and 1166 individuals of 71 species with point counts; from these, eleven species were exclusively recorded by point counts and two species exclusively by sound recordings. The two field methods had similar performances (SC of 0.98 vs 0.99 for sound recordings vs point counts, respectively), while SC was 0.99 for each vegetation type. SC per season ranged from 0.94 to 0.98 (Table S4).
Alpha diversity—Species diversity was consistently higher in OW than in TDF (Figure 2A), as was the case of functional diversity (Figure 2B). In the case of phylogenetic diversity, OW had a higher value only when q = 0, being equal to TDF when q = 1 and q = 2 (Figure 2C). In terms of IF0,2, we found that both OW and TDF had similar values (2.92 vs 2.96) when species diversity was considered, while OW showed higher dominance than TDF in terms of both functional (2.17 vs 1.86) and phylogenetic (3.85 vs 3.25) diversities.

Figure 2. Alpha diversity profiles of the two studied vegetational types. (A) Species diversity, (B) Functional diversity, (C) Phylogenetic diversity. Ribbons depict 95% confidence intervals. TDF: Tropical Dry Forest, OW: Oak Woodland.
In a seasonal basis, regarding the three types of diversity we found that the OW_M1 was consistently more diverse than TDF regardless of which specific diversity metric was used (Figure S1A-C). In the breeding season (B), both vegetation types were equally diverse across orders and diversity types (Figure S2A-C) excepting functional and phylogenetic diversity of order 0 (OW had slightly higher values than TDF). In M2, OW was consistently more diverse than TDF at all orders and diversity types considered (Figure S3A-C), following similar patterns as M1.
Beta diversity and species composition—On a yearly basis, species beta diversity was moderate, being highest when q = 0 (0 D = 1.4) and decreasing when q = 1 and q = 2 (1 D = 1.29; 2 D = 1.27) (Figure 3). In terms of phylogenetic beta diversity, it was low at q = 0 (0 D = 1.2) and decreased as q did (1 D = 1.05; 2 D = 1.01), mirroring the pattern of the species diversity. Functional beta diversity was low at q = 0 (0 D = 1.32), increasing at q = 1 and q = 2 (1 D = 1.67; 2 D = 1.63) (Figure 3). Regarding seasonal beta diversity, both species and phylogenetic diversities showed similar trends across seasons (decreasing values as the diversity order did) and functional diversity increased as q did (Figs. S1D, S2D, S3D).

Figure 3. Profiles of the comparison of beta species, functional and phylogenetic diversities between Tropical Dry Forest and Oak Woodland.
The gap statistic analysis indicated the optimal delimitation of 14 functional groups (Figure S4). The largest group was constituted by resident foliage-gleaner frugivore-insectivore species, which were more abundant in the OW than in TDF (Figure S5). The ground-granivores and migratory insectivore foliage-gleaners were also more abundant in OW, but large tree-frugivores insectivores and large frugivores, the latter represented by only one species (Plain Chachalaca Ortalis vetula), were more abundant in TDF. Fisher exact test was significant (p < 0.001). In the TDF, six functional groups did not differ in observed abundances with respect to expected, while in OW, only two functional groups did not differ in abundances with respect to expected values (Table S5).
In terms of community structure and composition, the dominant species in TDF were Streaked-backed Oriole (Icterus pustulatus), Golden-fronted Woodpecker (Melanerpes aurifrons), Ash-throated Flycatcher (Myiarchus cinerascens) and White-winged Dove (Zenaida asiatica), while the dominant species in OW were White-tipped Dove (Leptotila verrauxi), White-winged Dove, Bridled Sparrow (Peucaea mystacalis) and Russet-naped Wren (Campylorhynchus humilis) (Figure 4). Rare species in TDF included Black-vented Oriole (Icterus wagleri), Boat-billed Flycatcher (Megarhynchus pitangua), Bronzed Cowbird (Molothrus aeneus) and Sumichrast Sparrow (Peucaea sumichrasti), while in the OW the rarest species were Pileated Flycatcher (Xenotriccus mexicanus), Dusky Flycatcher (Phaeoptila sordida), Ferruginous Pigmy-Owl (Glaucidium brasilianum) and Laughing Falcon (Herpetotheres cachinnans) (Figure 4).

Figure 4. Rank abundance plots of the two studied bird communities. Grey line: Tropical Dry Forest; black line: Oak Woodland. The four commonest and the five rarest are depicted for each community. Species keys can be seen in Table 1.
Bird distribution—The chi-squared test of the relationships between species (rows) and sites per season (columns) was significant (X = 1998.6, DF = 360, p < 0.001). The first two axes of the correspondence analysis represented 41.8% (eigenvalue 0.406) and 23.4% (eigenvalue 0.228) of the total variance, respectively. The samples of TDF were located at the positive end of the axis 1, while the samples of OW were located at the opposite extreme. The greatest contribution to axis 1 was made by TDF site in all the seasons and OW in the breeding season (Figure S6). Species with the highest contribution to axis 1 were I. pustulatus, Myiozetetes similis, P. leclancherii, C. formosus and M. aurifrons, which were located at the positive end of the axis. On the other hand, the species with the highest contribution to the axis and located at its negative extreme were P. mystacalis, Aimophila rufescens, Banded Wren (Thryophilus pleurostictus), Woodhouse’s Scrub-Jay (Aphelocoma woodhouseii) and Gray-breasted Woodpecker (Melanerpes hypopolius).
Axis 2 was defined by the differences in bird distribution between TDF during the breeding season (at the negative extreme of the axis) and the two migratory seasons (at the positive extreme of the axis). Species with the highest contribution to the positive extreme of the axis were M. cinerascens, Blue-gray Gnatcatcher (Polioptila caerulea), Yellow-green Vireo (Vireo flavoviridis), C. formosus and Z. asiatica. Species with the highest contribution to the negative extreme of the axis were Rufous-backed Thrush (Turdus rufopalliatus), Sulphur-bellied Flycatcher (Myiodynastes luteiventris), Russet-crowned Motmot (Momotus mexicanus), P. leclancherii and Plain-capped Starthroat (Heliomaster constantii) (Figure S6).
Finally, the chi-squared test of the relationships between distributional groups and vegetation types was significant (X = 66.30, DF = 5, p < 0.001). Neotropical and widespread species were positively associated to the TDF, while endemic and Nearctic species were so in the OW (Figure S7). Two distributional groups did not differ from expected abundances in TDF, while only one distributional group did not do so in OW (Table S6).
Discussion
Contrary to our expectations of a monotonic decrease of diversity with altitude, we consistently found higher alpha diversity values in OW than in TDF excepting phylogenetic 1D and 2D, where diversity was equal between vegetation types. Studies on the comparison between alpha bird diversity of TDF and OW are scarce. In Cuicatlán, north Oaxaca, a higher richness in TDF than in OW was found, although the study only included data from the migratory season (Peterson et al. Reference Peterson, Escalona-Segura, Zyskowski, Kluza and Hernández-Baños2003). In a study made in Guerrero, Mexico, a similar number of resident species (58 in TDF vs 61 in OW) was found at both vegetation types (Vázquez-Reyes et al. Reference Vázquez-Reyes, Arizmendi, Godínez-Álvarez and Navarro-Sigüenza2017). In terms of bird richness, our findings partially matched these previous results. Albeit the taxonomic diversity was higher in OW than in TDF after accounting for abundances. Contrary to expected, taxonomic diversity does not decrease monotonically with altitude (McCain Reference McCain2009), so either a mid-domain effect or a monotonic increase in diversity with altitude may be occurring (Montaño-Centellas et al. Reference Montaño-Centellas, Loiselle and McCain2021). This well-known relationship between species diversity and altitude (monotonic decrease) has been reported in a geographically close region with similar environmental conditions as well (Navarro-Sigüenza Reference Navarro-Sigüenza1992). Yet, contrasting hump-shaped patterns have been documented in more southerly latitudes (Araneda et al. Reference Araneda, Sielfeld, Bonacic and Ibarra2018).
In a seasonal basis, OW showed higher diversity values than TDF only during the migratory seasons. Considering that most migrants in our area are gleaner insectivores, as is the case of other OW in western Mexico (Hutto Reference Hutto1994; Corcuera Reference Corcuera2001; Corcuera and Zavala-Hurtado Reference Corcuera and Zavala-Hurtado2006), a higher number of individuals spend the non-breeding season at higher altitudes, maybe because of a high invertebrate abundance in OW (Greenberg and Bichier Reference Greenberg and Bichier2005). This reflects the importance of exploring the seasonal differences in diversity in these dry vegetation types (Gonçalves et al. Reference Gonçalves, Cerqueira, Brasil and Santos2017). The seasonal pattern of higher diversity in the OW exhibits the preference of most Nearctic-Neotropical migrant birds for wintering at mid altitudes.
We found a decreasing trend in both taxonomic and phylogenetic beta diversities as diversity order increased, and the reverse trend when functional diversity was assessed, partially matching our work hypothesis. The low phylogenetic beta diversity may be an indicative of species filtering, which is in accordance with the patterns observed in other dry regions in the tropics. For instance, Graham et al. (Reference Graham, Parra, Rahbek and McGuire2009) found evidence of phylogenetic clustering of the hummingbird communities in the seasonal dry slopes of the Andes, caused by the harsh conditions in the high and dry mountains which acted as an ecological filter. The low phylogenetic beta diversity indicates a close evolutionary relationship between the avifaunas of the two vegetational types. According to Becerra (Reference Becerra2005), the distribution of the NSDF is the result of successive events of isolation during the interglacial events. Such events gave origin to refugia, followed by periods of interconnection during the glacial periods (Castillo-Chora et al. Reference Castillo-Chora, Sánchez-González, Mastretta-Yanes, Prieto-Torres and Navarro-Sigüenza2021). Arguably, the low phylogenetic beta diversity between the studied vegetational types may be indicative of niche conservatism during the earliest stages of the NSDF (Chan et al. Reference Chan, Arroyo-Cabrales, Prieto-Torres and Sánchez-González2024). By contrast, the high functional beta diversity that we observed was caused by biotic interactions leading niche differentiation, in this case a few closely related species evolving to fill the available niches in the NSDF (Boyce et al. Reference Boyce, Shakya, Sheldon, Moyle and Martin2019; Arango et al. Reference Arango, Villalobos, Prieto-Torres and Guevara2021). Do Nascimento et al. (Reference do Nascimento, Correia, Ruiz-Esparza and Gouveia2018) documented differential responses of both functional and phylogenetic diversities along an environmental gradient in the Caatinga, with functional diversity depending on habitat heterogeneity and phylogenetic diversity associated to climate.
According to the seminal work of Janzen (Reference Janzen1967), beta diversity between altitudinal bands in tropical mountains is expected to be high due to the low species niche width, low dispersion rates and the constant climate across the year when compared to mountain systems in temperate latitudes. Both beta species and phylogenetic diversities between our vegetation types were low, contrary to the findings of Montaño-Centellas et al. (Reference Montaño-Centellas, Loiselle and McCain2021), who found that beta diversity was higher in the tropical highlands, something that we found only in the case of functional diversity. High bird beta diversity between altitudinal bands was recorded in a wet altitudinal gradient of Central America (Blake and Loiselle Reference Blake and Loiselle2000), and along a small altitude gradient of the dry forests in Western Mexico (Balvanera et al. Reference Balvanera, Lott, Segura, Siebe and Islas2002). Possibly, positive biotic interactions (e.g. between plants and birds) and not competition may play a major role in shaping a high functional beta diversity (Navarro-Sigüenza Reference Navarro-Sigüenza1992; Rahbek Reference Rahbek2005).
Regarding distributional patterns, we found bird species and distributional groups as exclusive of each vegetation type. It is known that the avifauna of the NSDFs in Southwestern Mexico is composed of a mixture of endemics, lowland and highland taxa (Prieto-Torres et al. Reference Prieto-Torres, Rojas-Soto, Bonaccorso, Santiago-Alarcon and Navarro-Sigüenza2019). In addition, the mountains of Mexico are comprised within the Mexican Transition Zone (Halffter and Morrone Reference Halffter and Morrone2017), biogeographic region inhabited by ecological assemblages of both Nearctic and Neotropical affinities plus restricted-range species. Our results show that Neotropical and widespread species (sensu Palomera-García et al. Reference Palomera-García, Santana and Amparán-Salido1994) are positively associated to the TDF (at lower altitude) while Nearctic and endemic species are also positively associated with the OW (at higher altitude).
Less than 10% of the total NSDFs are under official protection (Prieto-Torres et al. Reference Prieto-Torres, Nori, Rojas-Soto and Navarro-Sigüenza2021). Our data showed that although both species and phylogenetic beta diversity are low, functional beta diversity is high, suggesting a high complementarity between vegetation types that is important in terms of ecosystem services and ecological resilience (Winter et al. Reference Winter, Devictor and Schweiger2013). Although the advice of decreeing new natural protected areas in our study site might seem appropriate (Gillespie and Walter Reference Gillespie and Walter2001), the social, cultural and economic contexts suggest that other conservation and management schemes could fit better for the region, especially for dry forests (Castillo et al. Reference Castillo, Magaña, Pujadas, Martínez and Godínez2005). Around 75% of the land in Oaxaca is under social ownership, which implies a strong collective organization and control over the territories. These land tenure schemes allow the state to be one of the most biodiversity of the country, even when less than 10% of the Oaxaca’s surface is under official protection. This highlights the role of both peasant and indigenous communities in actively preserving the local biodiversity (Van Vleet et al. Reference Van Vleet, Bray and Durán2016). Currently, agro-industrial activities, transnational projects such as the Transisthmian Corridor, and the intensification of mining activities may jeopardize the local diversity in the short-term.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S0266467425100333.
Acknowledgements
We express our gratitude to the people of San Pedro Totolapam for allowing us to work in their municipality. Special thanks to L. Canseco, H. Fuentes, E. Mora, E. Cruz, and Angel for their assistance during the fieldwork. Thanks to the Xeno-canto project and contributors for allowing free access to a comprehensive set of bird recordings. Special thanks to four anonymous reviewers who critically revised and improved our manuscript. Fieldwork was done under permission of the Mexican government (SGPA/DGVS/01470/21). We acknowledge Universidad Autónoma Metropolitana for supporting the publication of this paper through the Transformative Agreements programme.
Financial support
The authors thank Don David Gold Mexico for funding this research´s fieldwork, and OSG thanks the Mexican Secretariat of Science and Technology (SECIHTI) for awarding him a postdoctoral grant (418758).