Urbanization is known to be a major driver in abundance and species richness of birds. However, how birds respond to urbanization in tropical cities is understudied in general and entirely absent from Myanmar. We conducted a study in and around Mandalay, a large city in central Myanmar to gather first data on birds and their response to urbanization. We selected four habitats with 10 sampling points each in November 2015. We made 1,536 observations of 68 bird species. The number of species and diversity significantly differed between the four localities. Mandalay Hill and Downtown Mandalay had the lowest number of species and diversity, whereas the University Campus and Paddy Fields had the highest. The highest number of observations was in Downtown Mandalay (1,003 counts) and the lowest on Mandalay Hill (103). Nonmetric multidimensional scaling ordination techniques showed that the four habitat types had significantly different bird species composition. Our results indicate a large effect of urbanization on species diversity, species richness, and species composition of birds.
Introduction
Urban development has rapidly increased worldwide and an estimated 70% of the world's human population will be living in cities by 2050 (UN-HABITAT, 2013). Urbanization alters natural habitats and is considered as a major cause of the extinction of native species (Czech, Krausman, & Devers, 2000; Marzluff, 2001). Birds are highly affected by urbanization (e.g., Beissinger & Osborne, 1982; Biamonte, Sandoval, Chacón, & Barrantes, 2011; Crooks, Suarez, & Bolger, 2004; Reale & Blair, 2005; Sorace & Gustin, 2010; Suarez-Rubio, Renner, & Leimgruber, 2011). In general, bird species diversity and richness decrease in urban areas compared with native habitats, whereas biomass and density increase (Chace & Walsh, 2006; McKinney, 2008; Suarez-Rubio et al., 2011). However, this pattern may not hold for all geographic regions (Saari et al., 2016). Species composition also changes with urban development. Urban assemblages are similar in species composition compared with nonurban assemblages, indicating that urbanization is a major cause of biotic homogenization (Clergeau, Croci, Jokimaki, Kaisanlahti-Jokimaki, & Dinetti, 2006; Filloy, Grosso, & Bellocq, 2015; McKinney, 2006). Bird communities are usually dominated by a few, often introduced species, and granivores, omnivores, and cavity-nesting species are favored (Beissinger & Osborne, 1982; Blair, 1996; Fernández-Juricic, 2001; Ortega-Alvarez & MacGregor-Fors, 2009). These patterns have been confirmed throughout the world for forest, desert scrub, and grassland habitats (Chace & Walsh, 2006). Nonetheless, diversity patterns and composition of urban bird communities depend on abiotic factors and ecological and evolutionary forces at play (e.g., species interactions, immigration, and natural selection) (Faeth, Bang, & Saari, 2011).
Although Southeast Asia has a rapid human population growth, the level of urbanization is fairly low compared with other regions (Jones, 2013). Nonetheless, urban areas have been rapidly increasing in the last decades (Jones, 2013). In 2010, around 2% of Southeast Asia's population lived in urban areas, twice the proportion in 1970 (Jones, 2013). However, there is a paucity of information on the impacts of urbanization on avian communities in Southeast Asia (Chace & Walsh, 2006; Magle, Hunt, Vernon, & Crooks, 2012) and Myanmar in particular. Myanmar harbors some of the world's most biodiversity-rich ecosystems (Mittermeier, Turner, Larsen, Brooks, & Gascon, 2011; Sodhi, Koh, Brook, & Ng, 2004) and has been recognized as a biodiversity hotspot given the high concentration of endemic species in this area (Mittermeier et al., 2011; Myers, Mittermeier, Mittermeier, da Fonseca, & Kent, 2000; Sodhi et al., 2004). However, little is known about the effects of urbanization on Myanmar's avifauna despite the rich diversity that might be under pressure from human population growth.
The aim of this study was to examine bird species richness, diversity, and relative abundance along an urban–rural gradient in and around Mandalay city in central Myanmar and to assess bird community composition in areas with different housing density. We expected that areas with high housing density had low number of bird species, low diversity, and high relative abundance, whereas areas with low housing density and higher proportion of trees had high number of species, high diversity, and low relative abundance, as well as a shift in species composition.
Methods
Study area
The study was conducted in and around Mandalay (21 °59′N, 96 °5′E), Mandalay Region, in central Myanmar (Figure 1). Mandalay is the second largest city of Myanmar with around 1.5 million inhabitants and a population density of 7,000 people km−2. Population of Mandalay increased from 960,000 to 1,200,000 between 1993 and 2003 and the city is projected to increase 10% by 2025 (United Nations, 2008). The city is located in the central dry zone of Myanmar, in which total annual precipitation is 915 mm on average. Mean annual temperature is 27.3℃ and ranges between 13℃ and 38℃. Elevation is 74 m a.s.l., except for Mandalay Hill, which is located at the outskirts of the city, where elevation reaches 223 m. The typical vegetation in Mandalay is dry deciduous tropical forests.
Figure 1.
Study area (star) in Myanmar (bold gray line) and locations of point counts in Mandalay (colored dots/circles). Each dots/circle is a point count locality in one of the four land use classes (names refer to the classes and major sites we use in this study); black lines are township borders in Mandalay.
We selected four habitat types based on differences in housing density (Figure 1). Downtown Mandalay (DM) is mainly located at the western part of the Royal Palace, is highly developed ( > 70% impervious surface), and has a very high housing and human population density. Trees, mostly native species, are limited to few rather small patches and along streets. Mandalay University Campus (UC) is located south of downtown. It is moderately developed (30–40%), has many large trees, and partially dense understory. Some parts are covered with grass or bare soils. Mandalay Hill (MH) is located northeast of the city, is sparsely developed ( < 30%), and is covered with forests that consist mainly of scrubs and thorn bushes of up to 15 m height. Buildings consist of few pagodas/stupas and the stairways. In addition, there are few houses, and inhabitants living on the Hill are restricted to the monks. Paddy Fields (PF) at the eastern fringes of Mandalay near the village of Yay Kyi (around 300 m distance northeast) are rural agricultural areas with no houses and some larger trees. The agricultural area is characterized by a patchwork of rice Paddy Fields and Eugenia plantations, although rice is the dominant crop in the area.
At each habitat type, we selected 10 sampling points (i.e., 40 points in total) and kept a minimum distance of 250 m in-between point counts to ensure independence and considerably decrease the likelihood of counting the same bird twice. In Mandalay Hill, we placed the points along the main road to the pagoda and stairways to minimize interference with religious activities. In the other three places, we kept on roads or walking paths to have similar observational conditions and reduce bias by diversification of microhabitats. The characteristics of habitat, housing density, and general structure within each of the four areas are homogenous as to reduce any other potential effects on the bird species composition and relative abundance during analysis.
Bird surveys
We surveyed birds at the four habitat types from 20 to 24 November 2015. We used fixed-radius point counts with a 50-m radius. At each sampling point, we surveyed all birds heard or seen during 10 min from 6.30 a.m. to 10.00 a.m. We covered the same time span in each of the four habitat types to guarantee the same effort and procedure in all four areas. We recorded all birds and number of observations and noted weather, time, distance to observer, sex, and age as precise as possible. Each audiovisually recorded bird was likely an individual, but we used the term observation instead because we cannot certainty exclude double counts of one individual, because in few instances records may be impossible to tell apart. By using 10-min intervals, we decreased the probability of double counting and increased the probability of recording all species at any given sampling point during the 10-min intervals (Bibby, 2000). We also recorded all birds in-between point counts but did not use these for statistical analysis.
Data analysis
To evaluate differences between habitat types and number of species, we performed a generalized linear model (GLM) with a Poisson error distribution and checked for overdispersion. When overdispersion was detected, we corrected the standard errors using a quasi-GLM model (Zuur, Ieno, Walker, Saveliev, & Smith, 2009). Differences between habitat types and Shannon diversity index were assessed using analysis of variance (ANOVA) after verifying for homogeneity of variances (Fligner test) and normality (Bartlett test). We did a Tukey test for a posteriori comparison of diversity of the four habitats. We also assessed differences between habitats and observations (i.e., relative abundance) using a GLM with a Poisson error distribution. We performed this latter analysis both including and excluding the Rock Pigeon, because Rock Pigeons accounted for 78.9% of all observations in Downtown Mandalay. All analyses were done in R version 3.1.1 (R Core Team, 2014) and an alpha level of 0.05. We also assessed the abundance-based sampling coverage (C.hat, S.hat), calculated with Hsieh, Ma, & Chao (2013) with 500 bootstraps, per habitat type.
We assessed differences in bird community composition among habitat types using nonmetric multidimensional scaling (NMS; McCune & Grace, 2002). Relative abundance were log-transformed and then we searched for outliers among habitat types and bird species, using Sørensen (Bray–Curtis) distance with a cutoff of two standard deviations (McCune & Grace, 2002). We ran NMS using Sørensen distance, with a first approximation run of 6D stepping down to 1D solution, starting 20 runs from a random configuration and 250 iterations. We selected 2D as the final solution, using the starting configuration that worked best, and one real-run as suggested by McCune and Grace (2002). To evaluate significant differences among habitat types and bird species composition, we performed a distance-based permutational multivariate ANOVA (PerMANOVA) (Anderson, 2001) using 4,999 number of randomizations and using Sørensen distance. The outlier analysis, NMS, and PerMANOVA were performed using PC-ORD 6 for Windows (McCune & Mefford, 2011).
Results
During our survey, we recorded 1,536 in 68 bird species, with Rock Pigeon, House Sparrow, Streak-eared Bulbul, and House Crow the species with the overall most encounters (Appendix A). Our sampling coverage per habitat type (C.hat) varied from 0.923 in Mandalay Hill to 0.994 in Downtown Mandalay and is overall very high approaching 1.000. Number of species significantly differed between the four habitat types (GLM: Residual deviance = 32.480, df = 36, p < .001), also for estimated species richness per habitat type (Chao1: DM = 25.6, UC = 48.6, MH = 34.6, PF = 67.1). Mandalay Hill and Downtown Mandalay had the lowest number of species, whereas the University Campus and Paddy Fields had the highest (Figure 2). Bird diversity was also significantly different among habitats (ANOVA: F3 = 11.7, p < .001; Figure 2). Downtown Mandalay had the lowest bird diversity and was significantly different from University Campus (p = .002) and Paddy Fields (p < .001), which both had the highest diversity. Also Paddy Fields had high bird diversity compared with Mandalay Hill (p =.018). Observations were different among the four habitat types (including Rock Pigeon: GLM Residual deviance =3,450, df = 328, p < .001; excluding Rock Pigeon: GLM Residual deviance = 469.02, df = 308, p < .001). The highest number of observations was in Downtown Mandalay (1,003 counts; Figure 3) and the lowest on Mandalay Hill (103).
Figure 2.
(a) Number of species among four habitat types in Mandalay, Myanmar, and (b) bird species diversity among four habitat types in Mandalay, Myanmar: Downtown Mandalay, University Campus, Mandalay Hill, and Paddy Fields.
Figure 3.
Variability in number of observations per locality (a) including and (b) excluding the Rock Pigeon in Mandalay, Myanmar.
Neither habitat types nor bird species were identified as outliers. NMS analysis produced a final optimum 2D ordination space, which represented 75% of the variance in the original species data (R2 NMS1 = 0.43, NMS2 =0.32, stress = 0.15). The NMS ordination showed that habitat types had different bird species composition (Figure 4). This was supported by the PerMANOVA, which showed a significant difference among habitat types (p < .001). Pairwise comparisons also indicated significant differences on species composition for the different habitats (Appendix B). An examination of the relative abundance of some species showed that some birds such as Eastern Stonechat tend to prefer Paddy Fields, others such as Yellow-browed Warbler may favor Mandalay Hill, and others such as Common Myna occurred more often in University Campus and Downtown Mandalay (Figure 5).
Discussion
Our results indicate distinct responses by the bird communities towards housing densities and development. We found that urbanization has significant effects on bird species richness (observed and estimated), diversity, and abundance. With the data we collected, we can confirm our expectations that areas with high developed areas (particularly Downtown Mandalay) have low number of bird species, low diversity, but high relative abundance, whereas low developed areas and areas with higher portion of trees (particularly University Campus) had high number of species, high diversity, and low relative abundance. Our results contradict at least partially other studies; for example, others have found that bird richness is highest in residential areas with yards composed by trees, shrubs, and lawn (Blair, 1996; Leveau & Leveau, 2005; Marzluff, 2008). Our results support previous studies in which higher abundance of few species such as sparrows are found in areas with high housing density, while at the same time generalist species that feed on anything available and breed mainly in cavities occur (Suarez-Rubio et al., 2011; Suarez-Rubio & Thomlinson, 2009).
We were surprised by the relatively large number of bird species in the paddy fields, because agriculture usually has lower numbers of birds than, for example, forested sites (O'Connell, Jackson, & Brooks, 2000). But in our case, the close proximity of many bushes and hedge rows diversified the otherwise monotonous landscape, which could explain the high diversity. In contrast, we found very low numbers of individuals and relatively few species on Mandalay Hill, an area with diverse habitat structure with interspersed pagodas. Perhaps, the daily high activity of humans might drive the bird individuals into areas with less activity and therefore have not been recorded in our surveys.
Besides the change in species numbers and diversity, we found a significant shift in species composition. In high developed areas (Downtown Mandalay), we found more species that are commonly associated with human development such as Rock Pigeon or Sparrows. Similar trends have been found for urban areas where few generalist species increase in numbers and are present while absent in forested areas for example (Blair, 1996). In addition, significant species composition changes have been found along other trends of human development such as exurban development (Suarez-Rubio et al., 2011).
As in other cities of the world, the Rock Pigeon and the House Sparrow were very frequent in Downtown Mandalay, but interestingly other sensitive species such as Streak-eared Bulbul can still be found in the city, probably due to remaining forest fragments or parks within the city. Nonetheless, our results indicate that urbanization has an effect on the bird community although the level of urbanization in Mandalay is fairly low compared with other Southeast Asian cities.
Implications for conservation
We provided preliminary information on the bird community of Mandalay, Myanmar and showed the effects of development (housing) on the species composition, diversity, and abundance. Although we have a small sample size (but very high estimated sampling coverage with over 92%) and have considered only a short time period, several species tend to avoid downtown areas such as Mandalay. To our knowledge, this is the first time that the effects of urbanization are being documented for Myanmar—a notoriously understudied country—and indicates its congruence with studies in other tropical and temperate regions (Sorace & Gustin, 2010; Suarez-Rubio et al., 2011; Suarez-Rubio & Thomlinson, 2009). Thus, continued monitoring of the bird community in cities of Myanmar, particularly Mandalay, would provide further insights on the effects of urbanization on birds, especially given the projected population growth in the city. In addition, remnant habitat and diverse vegetation structure within the city such as Mandalay University Campus should be preserved to aid the conservation of birds in the city.
Acknowledgments
The work originated from a training workshop on scientifically based methods for students and NGO members from Myanmar. The authors thank Rufford Foundation ( http://www.rufford.org/) for the kind support. The University of Mandalay hosted the training workshop and facilitated the work. Dr. Paul J. J. Bates (Harrison Institute) helped during different stages of the work to implement the workshop and gather field data. The funding organization and the host had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The authors thank Rufford Foundation ( http://www.rufford.org/) for funding the training workshop (Renner #17058-1).
