Introduction

Urbanization is perceived as a major driving force behind habitat fragmentation and conversion, but effects of urbanization on biodiversity loss is only recently being investigated at larger detail (Blair 1996; Blair 2001a, b; Blair and Launer 1997; Hunter 2007; Kowarik 1995; Marzluff 2001; McKinney 2002; Pauchard et al. 2006). Expansion of urban areas associated with mushrooming of urban sprawls and smaller suburban habitations has been shown to have serious bearing on the local species diversities increasing the local extinction rates (McKinney 2002, Marzluff 2001) and through replacements of native species (Kowarik 1995, Blair 1996, Blair and Launer 1997, Blair 2001a, b). Many of these changes are also said to be irreversible (Pauchard et al. 2006). From another perspective, analyses of the status of biodiversity along urbanization gradients would also provide useful information on species response to structural changes in complex landscape mosaics both at population as well as community levels (Bolger et al. 1997; Crooks et al. 2004; Germaine et al. 1998; Marzluff et al. 1998).

Large parks and reserves in urban areas may support high species diversity because these protected areas are the habitat “fragments” of highly diverse ecosystems (Schaefer 1994). Some studies of birds in urban areas have focused on species richness within a patch of native habitat (e.g., a large park) in relation to patch size and the degree of isolation from other areas of habitat (Tilghman 1987, Diamond 1988, Soulé et al. 1988). A few studies have examined the effects of surrounding urbanization on bird species richness in park and riparian fragments (e.g., Saab 1999, Sodhi et al. 1999, Mörtberg 2001, and Er 2002).

The occupancy of poor-quality or marginal areas by resident birds could be influenced by the surrounding landscape (regional parks vs. high-density housing) because the surrounding habitats may act as resource areas for residential birds and as “source” areas for dispersing birds. (Melles et al. 2003)

Urban habitats are altered from their original state, but do offer foraging and cover resources and are thus not always hostile or unsuitable for all species. In a complex landscape mosaic, the presence and detection of individual birds is expected to vary with land use and with the overall composition of the landscape (Trzcinski et al. 1999, Austen et al. 2001, Fahrig 2001).

Kolkata (earlier Calcutta) is the 7th biggest city in India in terms of area and population. By 2011 its population is predicted to grow up to 16.5 million which is a 22 % leap from 13.6 million in 2001. Housing densities are also increasing around the city as areas in the city outskirts are transforming fast to accommodate the bulging population. Expansion of Kolkata city into the adjoining rural landscapes, mostly through unplanned urbanizations at various scales, has created sharp urbanization gradients and has thrown open conservation issues that need to be addressed urgently.

The aim of this study is to analyze the avian community along urbanization gradients in selected surrounding areas of Kolkata. We attempted to see whether and how the bird communities are affected by the intensity of urbanization and if there is any change in the functional groups across the gradients. We made the following hypotheses:

  1. 1.

    Urban areas will have less number of species compared to the rural and Rural 3 areas.

  2. 2.

    Urban and rural areas will have different bird community structure.

  3. 3.

    Bird community along the gradient will show a nested pattern.

  4. 4.

    Bird species diversity will be determined by the extent of built up area along the gradient.

Methods

Study site

Three urban sites characterized by high house lot density, 1 peri urban site that was a recreational area on the outskirt of an urban settlement and 2 rural sites characterized by sparse house lots and more open space were chosen for the study (Table 1);

  1. 1.

    Rural 1(Barbaria)—A rural patch within 3 kms from a densely populated urban conglomeration with few dispersed houses and cultivation fields.

  2. 2.

    Rural 2 (Burul)—A village located 35 kms from the southern fringe of Kolkata metropolitan area. Characterised by low house lot density and had open cultivation fields with patches of thick vgeteation.

  3. 3.

    Rural 3 (Gandhighat)—A recreational area on a river bank area, adjacent to a densely populated urban conglomeration. This had good vegetation and low houselot density.

  4. 4.

    Urban 1 (Khardah)—A congested area with several housing complexes and industries. Sparse vegetation and almost no open space.

  5. 5.

    Urban 2 (Charkatala)—An urban area with high house lot density but with gardens attached to house lots. There were some water bodies in the area.

  6. 6.

    Urban 3 (Parnashree)—An area with a large water body surrounded by dense housing complexes. There were patches of open fields in this area.

Table 1 Characteristics of the sites chosen for bird diversity assessment and house lot densities in these sites
Full size table

House lot densities (housing units per square km) and extent of open areas (% of open space per square km) for these locations were calculated from Google Earth image (Fig. 1).

Fig. 1
figure 1

Landscape and Google Earth images of the study sites. a) Rural 1, b) Rural 2, c) Rural 3, d) Urban 1, e) Urban 2 and f) Urban 3

Full size image

Bird diversity

Relative abundances of individual bird species were estimated at 6 selected sites, by point count method. A point count is a count undertaken from a fixed location for a fixed time. It can be undertaken at any time of year, and is not restricted to the breeding season. Point counts can be used to provide estimates of the relative abundance of each species or, if coupled with distance estimation, can yield absolute densities, too (Buckland et al. 2001).

At each sites, four transect points were selected randomly within 1 sq.km. The distance between two transect points were approximately 150–200 m. The time spent at each transect was 15 min. The sampling was done twice a day (morning 6 am onwards & evening 4 pm onwards), once in a month for consecutive 3 months (February 2009 to April 2009).

Species falling within last quartile of the species abundance distribution were considered rare and the rest were considered common (Davidar et al. 1996). Uniqueness of a site was calculated as % of unique species of the total number of species in a site (Savitha et al. 2008).

Functional diversity

The bird species were categorized into six foraging groups. The categories were a) frugivore, b) insectivore, c) carnivore, d) granivore, e) nectarivore and f) omnivore. Omnivores were species that feed upon more than two types of food items. Some species e.g. Chestnut tailed Starling, Golden Oriole, Indian Koel, Orange Headed Thrush, Rufous Treepie, Red Whiskered Bulbul and Red-vented bulbul were considered both as frugivore and insectivore. White breasted water hen was considered both as insectivore and carnivore.

Weighted averages of species abundances were calculated for different functional groups. This was done by weighting the abundances of different functional groups with the house lot density in each area and then dividing the term with total abundance of all species in that functional group (Jongman et al. 1995; Fraterrigo and Wiens 2005).

Nestedness

Nestedness among the six sampling sites representing urban, peri urban and rural sites was calculated using Nestcalc (Atmar and Patterson 1993). Nestedness occurs when species recorded from a fragment is a subset of a larger (more species rich) fragment. Values of Nc and C were calculated and statistical significances estimated from species presence absence matrix using Nestedness Calculator (Wright and Reeves 1992). Nc is a count of number of times a species presence in a fragment correctly predicts its presence in a richer or larger fragment. C is a measure of the relative degree of nestedness that is not affected by matrix size (Wright and Reeves 1992).

Results

A total of 48 bird species were recorded from 2858 individuals pulled across sites. The Shannon-Weiners’ diversity index (H′) was highest in the Rural 2 (Burul) site and lowest in Urban 3 (Charkatala) site (Table 2). There was no significant difference in individual count between urban and rural areas.

Table 2 Bird species richness and Shannon-Weiners’ diversity index for various study sites
Full size table

Rural sites and the peri urban site had higher number of rare species compared to the urban sites (Fig. 2). Rural sites also had higher uniqueness compared to urban sites. Two of the urban sites (Urban 1-Khardah and Urban 2-Charkatala) did not have any unique or rare species. Thirty five percent of the species were unique to the rural and Rural 3 areas and were urban avoiders.

Fig. 2
figure 2

Distribution of common, rare (species in the last quartile of the species abundance curve) and unique species (a species found only in a particular site) across different sites

Full size image

There was a highly significant correlation between house lot density and bird species diversity (Spearman rank correlation), r = − 0.886, p < 0.01). Regression analysis shows a linear fit between house density and bird species diversity (Shannon Weiner’s diversity index) at different sites (Fig. 3). The best fit regression model was y = −0.0297x + 17.17, R2 = 0.51.

Fig. 3
figure 3

Best fit regression model showing relationship between house density (houses per sq.km) and bird species diversity (Shannon diversity index)

Full size image

Cluster analysis clearly groups Urban sites (Khardah, Charkatala and Parnashree) rural sites (Barbaria, Burul and Ganndhighat) as separate clusters (Fig. 3). Urban sites 1 and 2 (Khardah and Charkatala) were most similar in terms of bird community structure and this cluster had a relatively distant similarity with Urban site 3 (Parnashree). The two rural sites were closest to each other and although this cluster grouped with the rural 3 site, it was relatively at a larger distance Fig. 4.

Fig. 4
figure 4

Dendrogram (UPGMA, average linkage between groups) based on Euclidean distances between sites showing urban and rural sites clustering separately with respect to bird species diversity

Full size image

The rank-abundance curve was steeper in urban sites (averaged for all sampling sites) as compared to the rural sites (averaged across sampling sites) (Fig. 5). Rank abundance curve was gentler for the rural areas that also had more number of species.

Fig. 5
figure 5

Rank abundance curves for the urban and rural sites. Relative abundance values were averaged for rural and urban sites

Full size image

There was significant variance among various functional groups across different sampling locations (Two way ANOVA, F = 10.36, p < 0.0001). Rural sites had more number of frugivores (Mann Whitney U test, p < 0.05) and insectivores (Mann Whitney U test, p < 0.05) (Fig. 6). Necatrivores were found only in the rural areas (Fig. 6). Urban areas had more number of granivores (Mann Whitney U test, p < 0.05) Fig. 7.

Fig. 6
figure 6

Relative abundances of various functional groups across various sampling sites

Full size image
Fig. 7
figure 7

Maximally packed species by site matrix for 48 bird species across 6 sampling sites

Full size image

Weighted averages for granivores and omnivores were highest indicating the prevalence of these feeding groups in the areas with higher house lot densities (Fig. 8).

Fig. 8
figure 8

Weighted averages for bird abundance

Full size image

The bird fauna was found to be significantly nested (Fig. 6) with a fill of 43.9 % (Nc = 335, C = 0.54, p < 0.002) indicating that the species found in the less species rich urban areas are subsets of more species rich and larger rural areas.

Discussion

Our results were consistent with studies done previously that showed decreased bird species diversity due to urbanization (Blair, 1996, Blair 2001a, b; Blair and Launer 1997; Hunter 2007; Kowarik 1995; Marzluff 2001; McKinney 2002; Pauchard et al. 2006; Garaffa et al. 2009). Urbanization leads to changes in habitat structure and resource availability and that result into segregation of the avian community (Beissinger and Osborne, 1982; Fraterrigo and Wiens 2005). That the rural and Rural 3 sites recorded more number of species confirmed our first hypothesis.

The bird communities were clearly distinct between the rural and urban sites and this trend also confirmed our second hypothesis about segregation of the bird community along an urbanization gradient. Our analysis of the functional group matrix illustrates how the urban and rural-rural 3 bird communities segregated themselves along the resource base. There is a clear distinction between urban and rural sites in terms of various functional groups. There was preponderance of frugivores and insectivores in the rural areas which is attributed to presumably higher resource availability, e.g. trees and open areas in the rural sites as compared to urban areas. That the urban areas had more granivore birds corroborates earlier observations (Emlen, 1974; Lancaster and Rees, 1979; Beissinger and Osborne, 1982; Mills et al., 1989; Kluza et al., 2000; Fraterrigo and Wiens 2005; Chace and Walsh, 2006). Expectedly again, weighted average for granivores and omnivores abundances were found to be more which is because of their increased abundances in the urban areas with more house lot densities. Fraterrigo and Wiens (2005) also found the same pattern. Contrary to our expectation of the urban areas having more omnivore (generalist species), there was but no significant difference between the rural and urban areas in terms of omnivore species. Earlier studies had recorded more generalist species in urban areas (Boren et al. 1999). Higher abundance of house, crows, which are essentially omnivorous, in our rural and Rural 3 areas would have contributed to this non difference.

Rural communities were more evenly distributed as compared to urban areas which had high dominance of omnivorous species like house crows and common myna. Chiari et al. (2009) proposed that more species rich communities have larger populations and hence have low local extinction rates. Our study however does not corroborate this finding as there was no significant difference in individual counts between the urban and rural communities.

The communities in the urban areas were a nested subset of the larger species pool in the rural areas. Nestedness is an important measurement that has conservation implications and has been observed across many taxa. Platt and Lill (2006) in an important study had shown nestedness of urban avian communities within a larger set recorded in urban parks. This signifies the sensitivity avian community to variation in resource base. Higher percentage of unique or rarer species in rural areas also corroborates such response of the bird community to variation in resource base. Fernandez-Juricic (2002) had shown how human disturbance can create nestedness in bird community structure.

Our hypothesis that extent of built up area will determine the bird species distribution was corroborated by the high correlation between the house lot density and bird species diversity. Dinicola (1990) had shown earlier that houselot density is a good measure of imperviousness of ground substratum and can be considered as an important parameter for habitat characterization. Melles et al. (2003) had found a significant relationship between the house lot density and bird species richness.

A study along a larger scale with more habitat parameters will throw more light on the responses of bird communities to various levels of resource availability due to habitat disturbances. This study also shows the need for creating and preserving more wilderness areas around the urban nucleus. Restoration of city parks would definitely go a long way in conserving the avian community in and around the Kolkata metropolitan area.