Biological Invasions

, Volume 11, Issue 4, pp 951–958

Spatial distribution of aquatic birds in Anavilundawa Ramsar wetland sanctuary in Sri Lanka


  • A. M. Gunaratne
    • Department of Aquaculture and Fisheries, Faculty of Livestock Fisheries and NutritionWayamba University of Sri Lanka
    • Department of Aquaculture and Fisheries, Faculty of Livestock Fisheries and NutritionWayamba University of Sri Lanka
  • C. N. B. Bambaradeniya
    • IUCN – The World Conservation Union
Original Paper

DOI: 10.1007/s10530-008-9307-1

Cite this article as:
Gunaratne, A.M., Jayakody, S. & Bambaradeniya, C.N.B. Biol Invasions (2009) 11: 951. doi:10.1007/s10530-008-9307-1


Biotic and abiotic factors that influence the avian distribution in a dry zone wetland was investigated by studying the distribution of Asian Openbill (Anastomus oscitans), Cotton Pygmy-goose (Nettapus coromandelianus) and Pheasant-tailed Jacana (Hydrophasianus chirurgus) in Anavilundawa Ramsar sanctuary in Sri Lanka in 2006. Their distribution was recorded in Anavilundawa, Suruwila and Maiyawa reservoirs, their catchments and respective paddy fields, through line transects. The floral cover of surface water was recorded by floating quadrates. Water lily (Nymphea spp.) was the dominant flora in Anavilundawa reservoir, invasive water hyacinth (Eichhornia crassipes) in Suruwila reservoir and a native lotus (Nelumbo nucifera) in Maiyawa reservoir, respectively. Anavilundawa area had a higher distribution of birds than in the other two areas. Among the three species observed, Asian Openbill distribution was mainly restricted to Anavilundawa whereas Pheasant-tailed Jacana was present in all three areas. Nesting of Asian Openbill occurred only in Anavilundawa while nesting of Pheasant-tailed Jacana was only observed in Maiyawa. Among the three species recorded, Cotton Pygmy-goose numbers were the least. Asian Openbill preferred dead trees and trees with no leaves for nesting and perching over live trees. It is concluded that the spread of invasives such as Water hyacinth and Salvinia (Salvinia molesta) forming mats over surface water have reduced the abundance of habitat specialists like Cotton Pygmy-goose.


Invasive floraEichhornia crassipesSalvinia molestaAsian OpenbillCotton Pygmy-goosePheasant-tailed JacanaAvian distributionNative invasives


Spatial distribution of a species indicates its ability to survive in a given habitat tolerating its physiochemical conditions, food availability and predator, competitor and parasite pressure (Begon et al. 2006). In many instances, spatial distribution of species show little change over time [ex. herons in England and Wales, (Marchant et al. 2004)]. However, when organisms extend their distribution on a geographical scale, it is likely that they are taking advantage of temporary, seasonal or permanent changes of climate or distribution of habitats that allow them to cross barriers that are normally closed (Hobday et al. 2006).

In Sri Lanka, anecdotal information is suggesting changes to the spatial distribution of several bird species. Amongst different habitats available for birds in Sri Lanka, wetlands are the focal point for a large number of both resident and migratory species (IUCN 2004). Similarly, wetlands are also the most susceptible habitats in terms of deterioration (Hails 1997).

The spatial distribution of birds in wetlands is likely be affected by many factors (Korschgen and Dahlgren 1992; Evans and Day 2002). In relation to spatial distribution of birds, two factors have been identified as critical; human disturbance (Skagen et al. 1991; Lafferty 2001; Fernandez-Juricic 2004) and introduction of exotic fauna and flora (Steadman and Martin 2003; Leyse et al. 2004).

It is known that 24% of the world’s most invasive plants are wetland species and they have the capacity to form monotypes (Murphy 1988; McConnachie et al. 2003) which alter habitat structure (Brockerhoff et al. 2003), lower the biodiversity (Smith et al. 2006) and change nutrient cycling and productivity (Siebert 1987; Kázmér 1990). Also, evidence on invasives affecting the bird diversity, is mounting. Blossey et al. (2001) reported habitat specialist birds such as marsh wrens (Cistothorus palustris) being affected by reduced habitats due to invasive forb Purple loosestrife (Lythrum salicaria). Similarly, Benoit and Askins (1999) reported significantly fewer species of birds in wetlands dominated by invasive Phragmites australis than in native cordgrass marshes.

In Sri Lanka, there are five species of recorded aquatic invasive flora and amongst them, Salvinia (Salvinia molesta) and Water hyacinth (Eichhornia crassipes) have spread to almost all wetlands ( Although the impacts of invasives have been extensively studied, their impacts on distribution of wetland birds have not been considered in many parts of the world. Therefore, spatial distributions of three species of wetland birds were studied in an ancient cascading system of five reservoirs in the dry zone of Sri Lanka.


The study was conducted in May 2005–2006 in the Anavilundawa Ramsar sanctuary (RAMSAR site No. 1038) which is located in North Western province of Sri Lanka (07°42′ N 079°49′ E) (Fig. 1).
Fig. 1

Anavilundawa RAMSAR site (Map based on

The elevation in the area is 0–1 m above sea level and the sanctuary covers an area of approximately 1,397 ha. The mean depth of the reservoirs varies between 3 and 4 m and mean area ranges between 12 and 50 ha, totally covering an area of approximately 200 ha, which helps to sustain about 412 ha of traditional paddy fields, ( The reservoirs are seasonal and they completely dry up during the dry season from August to November. The current study was restricted to three reservoirs, their catchments and their respective paddy fields, Anavilundawa (23 ha), Suruwila (12 ha) and Maiyawa (37 ha) (Fig. 1). This wetland is the final destination for approximately 37 species of migratory birds and about 131 species of resident birds ( In the current study, three bird species were studied. The largest of the three species studied was the Asian Openbill (Anastomus oscitans); the smallest Sri Lankan stork that use swamps and marshes for nesting and breeding (Harrison 2005). The Cotton Pygmy-goose (Nettapus coromandelianus) is the smallest local duck and is an almost entirely aquatic bird. The Cotton Pygmy-goose population has declined after 1980s for the reason of loss of their habitats and it is now considered as a rare species (Ceylon Bird Club Reports 1986–2005). Pheasant-tailed Jacana (Hydrophasianus chirurgus) is a long-legged marsh bird sometimes commonly known as the lotus bird.

Data were collected from May 2005 to 2006. Stratified random sampling was employed to lay the transects in reservoirs, paddy fields and catchments. Please see Fig. 2 for details of lengths and number of transects. In each reservoir, three transects were marked using visual clues in the two banks. The first transect in each paddy field started from a random point next to the bank of each reservoir. In catchments only one 50 m transect was marked along the road, as the catchment was inaccessible throughout the study period due to inundation.
Fig. 2

A schematic representation of the experimental design (Experimental design elaborated only for one area. Same experimental design was applied in the other two areas)

Data on study species distribution along the transects were collected in the morning (0615–1100) and in the evening (1430–1900) using a binocular (10 * 50). Sampling intensity varied from 7 days a week to once a week depending on accessibility mainly due to rain. During the breeding season Asian Openbill distribution was estimated by colony counting method (Sutherland 1996). Floating frame quadrate method (1 × 1 m—divided into 100 blocks) was used to determine the density of invasive and native flora.

In addition, data on DO, pH, turbidity, water temperature, and salinity of each reservoir were also recorded. In the case of Asian Openbill the type and the condition of trees on which the birds were nesting were also recorded.

ANOVA was used to compare the spatial distribution of study species between three areas and distributions of study species in different habitats within different areas. Variations in adults and sub adults abundances and nesting site preferences were analysed by unpaired two sample t-test. Male: female ratios of Pheasant-tailed jacana in three reservoirs and the effect of observation time were also analysed by ANOVA. The mean number of study species seen in each transect were transformed into natural logarithms and general linear regression was used to interpret the variations in spatial distributions using other explanatory variables (Zar 1984) (Table 1).
Table 1

The variables recorded during the study and their categories


Explanatory variable




Rainy/not rainy


Starting time

Early morning (0615–0744)/Mid morning (0745–0914)/Late morning (0915–1100)/Early evening (1430–1559)/Mid evening (1600–1729)/Late evening (1730–1900)


Wind speed

Beufort scale (1–12)


Cloud cover (%)

<24/25–49/50–74, 75<


Cloud type







Disturbed (the transects near/along the roads and near the bathing points)/Less disturbed


Condition of the nesting trees

Live trees/Live trees without leaves or dead trees


Between the three areas studied, Anavilundawa area had an overall higher distribution of the observed species (\( \bar x = 37, \) se = ±6 birds) than in Maiyawa (\( \bar x = 17, \) se = ±2 birds) and in Suruwila areas (\( \bar x = 1 4, \) se = ±2 birds) (ANOVAdf = 2, 249, P = 0.000) (Fig. 3).
Fig. 3

Distribution of total birds in three sites (bars depicting the mean number of birds with SE)

The mean number of study bird species seen in Anavilundawa reservoir was higher (\( \bar x = 7 5, \) se = ±7 birds) than in Suruwila (\( \bar x = 2 5, \) se = ±2 birds) and Maiyawa reservoirs (\( \bar x = 3 1, \) se = ±2 birds) (ANOVAdf = 2, 105, P = 0.000) (Fig. 4). There was no difference in the distribution of birds in the paddy fields of three areas. However, the catchment of Anavilundawa site had a higher distribution of birds (\( \bar x = 7, \) se = ±1 birds) than the other two catchments; Suruwila (\( \bar x = 2, \) se = ±1 birds) and Maiyawa (\( \bar x = 3, \) se = ±1 birds) (ANOVAdf = 2, 33, P = 0.000).
Fig. 4

Distribution of total birds in three reservoirs (bars depicting the mean number of birds with SE)

Among the three species observed, Pheasant-tailed Jacana numbers were highest in Maiyawa reservoir (\( \bar x = 2 8, \) se = ±2 birds) compared other two reservoirs (ANOVAdf = 2, 107, P = 0.000). Asian Openbill had a restricted distribution and they were present only in Anavilundawa reservoir (\( \bar x = 4 9, \) se = ±5 birds) with Cotton Pygmy-goose found mainly in Suruwila reservoir (ANOVAdf = 2, 107, P = 0.001) (Fig. 5).
Fig. 5

Distribution of mean number of birds with SE of Asian Openbill (black bars), Cotton pygmy-goose (gray bars) and Pheasant-tailed Jacana (open bars) in Anavilundawa, Suruwila and Maiyawa areas

Nests of Asian Openbill were found only in the trees in Anavilundawa reservoir and there were more sub adults (\( \bar x = 2 5 6, \) se = ±1 birds) than adults (\( \bar x = 2 4 2, \) se = ±1 birds) in the population (ANOVAdf = 1, 503, P = 0.013). They preferred to perch and nest in dead trees and trees without leaves (\( \bar x = 9, \) se = ±1 nests) over live trees bearing leaves (\( \bar x = 3, \) se = ±1 nests) (ANOVAdf = 1, 13, P = 0.000). There were more male Pheasant-tailed Jacana (\( \bar x = 1 9, \) se = ±1 birds) in all the three reservoirs than the females (\( \bar x = 7, \) se = ±1 birds) (ANOVAdf = 1, 166, P = 0.000). However, the sub adults of Pheasant-tailed Jacana were only found in Maiyawa reservoir.

The highest number of birds were observed during the early morning hours (0615–0744) and late evening hours (1730–1900).

In Anavilundawa reservoir, water lily (Nymphea spp.) and invasive Salvinia molesta were the dominant flora with only 45% of the reservoir available as open water (Fig. 6). In Suruwila reservoir, the dominant species found were invasive Eichhornia crassipes and Nymphea spp. and only 17% of the reservoir was available as open water. In Maiyawa reservoir Nelumbo nucifera, a native lotus had 97% of floral coverage with only 3% of the reservoir available as open water (Fig. 6).
Fig. 6

Floral distribution in Anavilundawa (A) Suruwila (B) and Maiyawa (C) reservoirs


The results indicate that despite being connected and located close by, the three species were showing unique distributions in the three reservoirs. Birds being mobile, still showing affinities to certain habitats over others, could be an indication of certain environmental parameters either favouring or restricting their dispersal. The differences seen in distribution could be discussed in relation to vegetative characteristics of three reservoirs, specially the invasives; with regards to spatial distribution of Asian Openbill (large colonies but restricted to Anavilundawa), Cotton Pygmy-goose (few individuals mainly in Suruwila) and Pheasant-tailed Jacana (equally distributed over three reservoirs).

In this study, it was revealed that despite being a part of a cascade system, the water columns of the three reservoirs were invaded by different floral species. Maiyawa reservoir which is the last reservoir in the cascade system had the highest plant cover which was 97%. Interestingly, it was mainly lotus, a native species that has spread to the level of an invasive. It is already known that native species show invasiveness in response to environmental changes (Lynch and Saltonstall 2002). As shown in the map (Fig. 1.), Maiyawa reservoir being the last to receive water from the system could be the “sink” of this system. It is already very shallow due to accumulation of sediments, thereby all the environment changes could have resulted in native flora to spread like an invasive thereby affecting the use of habitats by avifauna.

The distribution of Pheasant-tailed Jacana in all three reservoirs in similar numbers may indicate their resistance to differences in three reservoirs and perhaps they are taking the advantage of the spread of aquatic plants as mats over surface water. They are known to be adapted to walk on floating vegetation (Harrison 2005). However, their breeding was restricted to one corner of Maiyawa reservoir which is relatively free from lotus and other floating vegetation. The breeding ground was observed to have thick broad bushes growing in the water. It is known that Pheasant-tailed Jacana prefers partially covered small piles of floating vegetation for their breeding (Harrison 2005). Unless attempts are taken to maintain the characteristics of breeding sites, there will be a threat to their future breeding.

In contrast to other reservoirs, Anavilundawa reservoir was relatively free from floating vegetation. It could be due to its strategic location in the middle of the cascade system where water constantly flushes through (Fig. 1). Therefore, it could be assumed that Anavilundawa reservoir has the advantage of being a “source” than a “sink”. This reflects the same phenomenon as seen in other systems where impacts being highest at low lying part of an aquatic system compared to a wetland with a slope enabling water flow (e.g. lower course of a river compared to an upper course; Zedler and Kercher 2004).

As indicated by results, Anavilundawa reservoir was selected over other reservoirs by Asian Openbill for their breeding and nesting. This could be due to the presence of large number of veteran trees within the reservoir itself, a character not seen in other reservoirs. Mass breeding can attract nest predators mainly water monitors (Varanus salvator) and rat snakes (Ptyas mucosus). However the surrounding water could defer the predators to some extent and as Shochat et al. (2005) observed for common grackle (Quiscalus quiscula), that could be an anti-predation adaptation. However, the impacts of mass nesting could lead to other problems such as increased accumulation of nutrients (Ligeza and Smal 2003). The signs of the dying back of veteran trees within Anavilundawa reservoir may indicate the impacts of high nesting pressure on the trees. It is important to note that breeding coincides with vegetative growth (i.e., in November to February after receiving rain from inter monsoon). However, we observed that most trees, where the nesting occurred, were leafless during the breeding season. Hence stunting and retarded growth and death of trees is a possibility due to mass nesting as recorded by Johnson and Beck (1988) and Lapointe et al. (2000).

The distribution of Cotton Pygmy-goose in three reservoirs is interesting. Anecdotal evidence indicates high number of Cotton Pygmy-gooses in this area in the past. However, the current study detected only one flock comprising of 5–7 individuals, restricted mainly to Suruwila and Anavilundawa reservoirs. The total avoidance of Maiyawa reservoir which is nearly completely covered by Lotus could be due to lack of their preferred habitats. Their low numbers in other two areas might also spread of Salvinia and Water hyacinth reducing the percentage of open water. The presence of less Cotton Pygmy-goose in Anavilundawa reservoir which had more open water than in Suruwila reservoir which has less open water perhaps indicates the influence of other factors on the distribution of Cotton Pygmy-goose. Floral compositions of these reservoirs are quite different in terms of invasives with more water hyacinth in Suruwila reservoir than in Anavilundawa reservoir. Hence, this study hints that not only the level of openess water but also the actual floral composition could affect a species. Similarly, Blossey et al. (2001) recorded the impacts of invasive Lathryum salicaria reducing the abundance of numerous insects and bird species that are habitat specialists such as marsh wrens. Therefore, evidence is mounting that invasives reduce the abundance of habitat specialists (see Zedler and Kercher 2004).

With the declaration of this sanctuary as an international Ramsar wetland, human presence has increased. It is therefore important to consider this aspect in future as there is a tendency for spread of invasives in wetlands with high hydrological disturbances (OldeVenterink et al. 2002).


Cotton Pygmy-goose which is a habitat specialist, preferring open water, was found in the least numbers in all areas, and preferred Suruwila reservoir to Anavilundawa reservoir, perhaps indicating the impacts of spread of invasives. Selection of only Anavilundawa for breeding by Openbills could be due to many reasons such as availability of preferred food or/and as a result of an antipredator strategy, as nesting trees were within the reservoir. High breeding pressure of Asian Openbill could have an effect on veteran trees as those used for nest construction were showing signs of die-back. As waders like Pheasant-tailed Jacana seem to be generally distributed despite the floral differences in reservoirs and occurrence of invasives, future management strategies for controlling invasives needs to be well planned, so that no birds are unnecessarily affected.


The authors thank Department of Wildlife Conservation, Sri Lanka for their permission to conduct field work at Anavilundawa Ramsar sanctuary. M. D. S. T de Croos, K. Wijenayake, G. Thilakeratne and C. Pathirage are acknowledged for assistance in the field. We appreciate the help of Anavilundawa temple and villagers. Wayamba University of Sri Lanka funded the study.

Copyright information

© Springer Science+Business Media B.V. 2008