Diversity and distribution of avifauna at Warathenna-Hakkinda Environmental Protection Area in Kandy , Sri Lanka

The present study was carried out in the recently established Warathenna-Hakkinda EPA in the Kandy District, Sri Lanka to investigate the avifaunal diversity and conservation threats. Sampling was conducted in two main habitat types: river islands and riverine forests. Both point count (10 m radius) and line transect (200 m long) methods were utilized and maximum of 30 minutes was used to sample the birds in each point. Food habit, niche type, endemism, abundance, and diversity indices (Shannon & Margalef) were calculated to compare the two habitats in this area. A total of 74 bird species belonging to 61 genera and 35 families were encountered from the study site. Among these, seven species are endemic: Sri Lanka Grey Hornbill, Sri Lanka Green Pigeon, Sri Lanka Wood Pigeon, Sri Lanka Hanging Parrot, Crimson-fronted Barbet, Yellow-fronted Barbet, and Sri Lanka Hill Mynah; while six were migratory: Green Sandpiper, Common Sandpiper, Indian Pitta, Asian Brown Flycatcher, Yellow Wagtail, and Grey Wagtail and the rest were non-endemic natives. Most of them were canopy and sub canopy dwellers associated with riverine forests and islands. According to the National Red List of Sri Lanka, three species, the Sri Lanka Wood Pigeon, Sri Lanka Hill Mynah, and Alpine Swift, are listed as threatened. Ardeidae, Alcedinidae, Columbidae, Accipitridae, and Apodidae were the most species rich families in this habitat. Out of the sampled species, 31% and 25% of the birds were carnivores and insectivores, respectively, while 5% were nectarivores. According to the avifauna, the riverine forests are more diverse (Shannon index H’= 2.55; Margalef’s index M= 10.92) than the river islands (H’= 2.29; M= 5.07) in this landscape. The variety of habitats along the Mahaweli River at Warathenna appears to aid in sustaining a rich bird community and this Environmental Protection Area will help create a safe haven for the birds.

ISSN 0974-7907 (Online); ISSN 0974-7893 (Print)  (Bossuyt et al. 2004). Sri Lanka, which is a tropical island located in the Indian Ocean between longitudes 79.65E and 81.88E and latitudes 5.90N and 9.86N experiences a year round tropical climate with average temperatures of 27-28 °C and a relatively constant day length (Wijesundara & de Silva 2005). The precipitation in the island is largely governed by the south-west and north-east monsoons (Ashton et al. 1997;Gunatilleke et al. 2008). Sri Lanka is renowned for the diversity and endemicity of its vertebrate fauna, including its diverse avifauna ( Warathenna-Hakkinda EPA was declared in 2017 after the findings of rare, endemic flora and fauna from this area. This EPA has an extent of 0.61 km 2 extending from the Halloluwa suspension bridge in the north to the Gatambe Bridge in the south, and bordered by the Peradeniya-Halloluwa-Katugastota Road on one side and the "Srimath Kuda Rathwaththe Mawatha", "Deveni Rajasinghe Mawatha", and the Colombo-Kandy Road on the other side of the river (Sri Lanka 2017: 4/81.VI). This EPA encompasses an approximately 2 km stretch of the Mahaweli River, all its river islands in this stretch referred to as the Hakkinda river islands, as well as the surrounding riverine forests and Kandyan home gardens. Only a handful of studies thus far have looked at the biodiversity in this area including the diversity of dragonflies and fish (Samarawickrama et al. 2012;Kumara & Samarawickrama 2017;Thilakarathne et al. 2019). The present study aimed to evaluate the species composition of the bird community in this EPA.

Description of the study site
Mahaweli river is the longest river in Sri Lanka. It originates from the central highlands and flows through the wet, intermediate and dry zones of the country (Fernando 1990 and Cryptocoryne (C. undulata, C. beckettii, C. walkeri, and C. parva) were seen in the shallow waters around the river islands. Furthermore, the exotic free floating J TT Eichhornia crassipes was seen in areas with slow moving water.
The sampling of avifauna was carried out in 17 river islands and 24 riverine forest patches. Sampling locations were marked using a portable global positioning system (GPS) (Magellan eXplorist 110) and sampling points were separated by a minimum distance of 150 m (Figure 1).

Survey of avian fauna along the declared EPA
Field sampling was carried out twice a week during the study period. Systematic random sampling method was used to record the birds from this EPA, and sampling was conducted along two main roads on either side of the Mahaweli River, the Peradeniya-Halloluwa-Katugastota Road (2.5 km) on the west and the "Srimath Kuda Rathwaththe Mawatha" and "Deveni Rajasinghe Mawatha" (3 km) on the east. Eighteen 200 m line  Taxa

Data analysis
Data on avifauna were analyzed for riverine forests and the river islands. The species richness, family richness, endemism, migratory percentages, relative abundance, and the diversity were calculated using R version 3.6.1 (R Core Team 2014). Pooled t-test was carried out to determine the significance of the two habitats under 95% confidant intervals using R version 3.6.1 (R Core Team 2014). The difference among different micro-habitats/strata was evaluated using one-way ANOVA using tukey comparison under 95% confidant limits in R studio version 3.6.1 (R Core Team 2014). The relative abundance was calculated as follows.
The rarefaction curve, endemism, family representation and national conservation status were plotted using SigmaPlot-10. Chao 2 and second order Jacknife 2 richness estimators were calculated in EstimateS software (Version 9.1.0) (Chao et al. 2005) to estimate the predicted species richness in the EPA. Species diversity was measured using Shannon diversity index. The proportional abundance of species was used to determine the Shannon diversity index (H') and the species richness was used to determine Margalef's diversity index using R version 3.6.1 (R Core Team 2014) for comparison of avifaunal diversity between riverine forests and river islands. Jaccard similarity index was calculated using R version 3.6.1 (R Core Team 2014) to estimate the degree of similarity between the two habitat types.
Where, p i = Relative abundance of a species, S= Number of species (species richness), and N= Total number of individuals Where, j= Number of species found in both sites, a= Number of species in site A, and b= Number of species in site B

RESULTS
A total of 666 individuals belonging to 74 bird species, 61 genera and 35 families were encountered during the study (Table 1, Image 1,2). The species based rarefaction curve, with lower and upper limits, did not reach the asymptote, which implied that more sampling effort was needed. According to the Chao 2 and Jackknife 2 estimates, the species richness could be 88 and 104, respectively ( Figure 3). However, majority of the sampled birds were nonendemic native resident birds. Since this is a riverine system, 24% of the encountered species were water associated birds while 76% were terrestrial birds. The most species-rich family was Ardeidae (herons and egrets; Image 1C,E,F) with seven species. Other than that, Alcedinidae (kingfishers -four species; Image 1A, B), Columbidae (pigeons -five species), Accipitridae (raptors -four species), and Apodidae (swifts -four species) were the most common families represented in the EPA. Of the recorded families, 43% were represented by only a single species. Although the species richness J TT was highest in Ardeidae, Columbidae, Alcedinidae, Accipitridae, and Apodiade families, the highest number of individuals in this ecosystem and hence the dominant families were Aredeidae (15% of relative abundance), Phalacrocoracidae (12% of relative abundance), and Sturnidae (11% of relative abundance; Table 1).
Out of the total number of species encountered, three species are listed under the categories of Vulnerable (VU) and Endangered (EN) according to the National Red List (MOE 2012). Of the remaining species, three are listed as Near Threatened (NT) and the rest of the 63 species are listed as Least Concern (LC). Of the recorded species, Alpine Swift is in the EN category while the Sri Lanka Wood Pigeon and Sri Lanka Mynah are listed as VU (MOE 2012). However, according to the global conservation status, the Sri Lanka Wood Pigeon is considered to be VU, Sri Lanka Hill Mynah as NT while the rest of the 72 species are considered to be LC. The LC category species were the most abundant (99% of relative abundance) in this ecosystem while VU species were the least abundant (0.3% of relative abundance; Table 1 and Figure 4).
According to the food habits, the encountered birds in the EPA included carnivores, omnivores, insectivores, piscivores, frugivores, nectarivores, and granivores. The species richness of the birds among the food habit categories was significantly different (F= 1.45; p <0.05). Of these, most species were categorized as carnivores (31%) and insectivores (25%) while nectarivores (5%) were the lowest in this riverine system. In addition, the highest numbers of individuals were recorded from carnivores and omnivores (35% and 26% of relative abundance, respectively) while the least number of individuals were recorded from nectarivores ( Figure 5).
Forests can be defined by four vertical forest strata namely, canopy, sub canopy, understory and the ground layer. Some of the bird species in the EPA were encountered from a single layer while others were

J TT
encountered from multiple layers such as canopy-sub canopy, canopy-understory, sub canopy-understory, sub canopy-ground and understory-ground. Most of the bird species were encountered from the canopy and sub canopy layers in the forest. According to the forest strata analysis, most of the birds were recorded from the sub canopy-understory layer (47% of total sampled species) while understory and understory-ground dwellers (15% of sampled species in each strata) were the least species rich forest strata in this ecosystem. The species richness among the nine combinations of the four main strata of the forest layer was significantly different (F= 2.97; p <0.05) ( Figure 6). The abundance of species was highest in sub canopy-understory (50%) and sub canopy-ground (26% relative abundance) while the least abundance was recorded from the understory (0.15% relative abundance).
During sampling of the river islands, 136 individuals belonging to 34 species in 20 families were recorded while riverine forest patches recorded 510 individuals belonging to 72 species in 34 families (Image 2B). All seven endemic species recorded during the study were observed in the riverine forests as well as six migratory species except Green Sandpiper and 60 non-endemic natives. On the other hand, three endemics (Sri Lanka Grey Hornbill, Yellow-fronted Barbet, and Sri Lanka Hill Mynah) and two migratory species (Common Sandpiper and Green Sandpiper) and 29 non-endemic natives were recorded from the river islands (Table 1). However, the species richness was not significantly different between the two habitats (t= -2.13, p >0.05). Green Sandpiper and Velvet-fronted Nuthatch Sitta frontalis were the only two species not recorded from riverine forest although they were present in river islands. Swift, Munia, Wagtail, and woodpecker species were absent in the river islands in the EPA (Table 1). However, the species abundance between the two habitats were significantly different (t= -3.89, p <0.05). Yet, the relative abundance of the individual species between two habitats were not significantly different. The highly abundant species in riverine forests were the House Crow Corvus splendens, Common Mynah Acridotheres tristis, Little Cormorant Microcarbo niger (Image 1D), Intermediate Egret Ardea intermedia (Image 1F), Yellow-billed Babbler Turdoides affinis (Image 2G), and Little Egret Egretta garzetta (Image 1E) (47, 38, 37, 30, 26, and 25 individuals, respectively) while Little Cormorant and Southern Hill Mynah Gracula religiosa (37 and 24 individuals, respectively) were highly abundant in river islands. The least common species in riverine forest were Greater Coucal Centropus sinensis, Sri Lanka Hill Mynah, Great Egret Ardea alba, Cinnamon Bittern Ixobrychus cinnamomeus, Yellow Wagtail, Asian Brown Flycatcher, Tickell's Blue Flycatcher Cyornis tickelliae, Crimson-backed Goldenback Chrysocolaptes stricklandi, Lesser Goldenback Dinopium benghalense, Indian Pitta, and Sri Lanka Hanging Parrot while Velvetfronted Nuthatch, Crested Serpent Eagle Spilornis cheela, Asian Koel Eudynamys scolopaceus (Image 2D), Whitethroated Kingfisher Halcyon smyrnensis (Image 1A), and Black-hooded Oriole Oriolus xanthornus were the least abundant species in river islands. Shikra Accipiter badius   Taxa | www.threatenedtaxa.org | 26 October 2021 | 13(12): 19689-19701 (Table 1). Both Shannon-Wiener and Margalef's diversity indices were highest in riverine forests (H ' = 2.55; M= 10.92) compared to the river islands (H ' = 2.29; M= 5.07). According to the Jaccard index the two habitats showed 43% similarity. According to the food habits of the sampled species, the highest species richness and abundance was recorded from carnivores in both river island and riverine forest habitats (82 individuals 15 species and 154 individuals 22 species, respectively; Table 1), whereas the lowest species richness and abundance was recorded from nectarivores in both these habitats (9 individuals 3 species and 31 individuals 4 species, respectively).

DISCUSSION
Riverine forest habitats along the Mahaweli River provide shelter for diverse animal communities (Sinha et al. 2019). Due to the high heterogeneity of the habitats in riverine ecosystem, it is home to a variety of mammals, birds, reptiles, amphibians, and many invertebrate communities. However, the bird diversity in the Warathenna-Hakkinda EPA has not been investigated to date. Data on avifauna in this EPA may be useful for future conservation efforts and ecotourism.
Rarefaction graphical representations are used to determine the sampling effort. The statistical significance of this curve is revealed using 95% upper and lower confidence limits (Colwell & Coddington 1994). This curve for the present study is close to leveling off (i.e., not completely leveled off); hence, it indicates the requirement of more sampling effort. However, according to the published literature this graphical representation does not directly reveal the total species richness (Magurran 2004). Therefore, more commonly used and statistically powerful Chao 2 and Jackknife 2 methods are used to estimate the predicted species richness in the EPA (Colwell & Coddington 1994;Colwell et al. 2012). Since this study was conducted during the daytime, the study possibly missed the nocturnal species and may account for the lower number of species in relation to the estimated numbers of 88 and 104 species in Chao 2 and Jackknife 2.
The 74 species of birds recorded from the EPA represents about 16% of the island's total bird species. This indicate that the species richness in this EPA is higher than those reported in many other areas around Kandy city limits (Karunaratna 1986;Wijesundara & Santiapillai 2001;Ellepola 2014;Weerakoon 2015;Hettiarachchi & Wijesundara 2017). However, the bird species richness in Warathenna-Hakkinda EPA is lower than what is recorded in the Gannoruwa Forest Reserve (84 species) which is located adjacent to the EPA with an extent of 2.1 km 2 (Rathnayake et al. 2016). The Warathenna-Hakkinda EPA is a mere 0.61 km 2 area with not only riverine forest and river islands but also Kandyan home gardens surrounding human habitation. Hence, this area is highly affected by anthropogenic influences compared to the Gannoruwa Forest Reserve.
This EPA harbors about 28% of the islands breeding residents (out of the 240 breeding residents in Sri Lanka) and 3% of the 213 migratory bird species in Sri Lanka. In addition, it represents 26% of endemic bird species of the island (MOE 2012;Gabadage et al. 2015), while only 11 endemic species have been recorded from the Gannoruwa forest (Rathnayake et al. 2016 The forest strata the birds inhabit correlate with their food habits. Most of the frugivores (leaf birds, barbets and hill mynahs) and insectivores are observed in the canopy and sub canopy forest layers while most of the piscivores are found on the ground layer and omnivores are encountered from sub canopy, understory and ground layers during the present survey. A large number of carnivores, are recorded from the EPA because this is a water associated habitat. But our findings were contrary to the study carried out in Udawatta-kele, which is a secondary forest in Kandy (Ellepola 2014; Weerakoon 2015) and the Knuckles range (Subasinghe & Sumanapala 2014), where they record a large number of insectivores and frugivores. A previous study carried out in the hill country of Sri Lanka documented that the insectivores (Sri Lanka Bush Warbler, Common Tailorbird), frugivores (Yellow-fronted Barbet), and nectarivores (Purple Sunbird) showed a marked

J TT
preference for primary montane forests while carnivores preferred primary montane forests and grassland-forest interface where they can easily spot their prey species (Wijesundara & Silva 2005).
Although there is a difference in both the diversity and distribution of birds between the two habitats studied in the EPA, a large number of common bird species (43% of sampled birds) are recorded from both the habitats. This is possibly because there is very little distinction between these two habitats in terms of floristic characteristics. Several studies have shown that the bird distribution and diversities highly depend on the heterogeneity of the habitats (Bambaradeniya et al. 2003;Wijesundara & Silva 2005;Karunarathna et al. 2011;Subasinghe & Sumanapala 2014;Weerakoon 2015;Rathnayake et al. 2016;Hettiarachchi & Wijesundara 2017;Chandrasiri et al. 2018). The higher bird diversity in the riverine forests in the EPA may be attributed to the availability of relatively more distinct habitats.
Since the Warathenna-Hakkinda EPA is surrounded by human habitation, a number of anthropogenic activities have a direct impact on the EPA. Construction and expansion of human habitation, logging, garbage, and sewage dumping can lead to the deterioration and loss of biodiversity in this EPA. Given the high diversity of flora and fauna, including birds, as indicated in this study, this area requires effective conservation planning, including proper landscape management strategies, strengthening of existing national environmental rules and policies, continuous monitoring programs, public education and awareness programs for safeguarding the biodiversity in this EPA.

CONCLUSIONS
A total of 74 bird species belonging to 61 genera and 35 families were encountered from the Warathennahakkinda EPA, which represent 16% of the island total bird species, 28% of the islands breeding residents, 3% of the migratory bird species and 26% of endemic birds of Sri Lanka. The variety of habitats provided by the riverine forests and river islands along the Mahaweli River of the EPA appears to aid in sustaining a rich bird community. However, since this EPA is located within the densely populated city of Kandy, it faces immense population pressures. Therefore, effective conservation planning including continuous monitoring programs, proper landscape managing strategies, strengthening of existing environmental policies, as well as educational and awareness programs are essential to minimize threats and safeguard the biodiversity in the EPA.