Diversity and temporal variation of the bird community in paddy fields of Kadhiramangalam, Tamil Nadu, India

Paddy, a major food crop of India, provides a variety of habitats in a short period of time and supports diverse organisms. Paddy fields also harbour many birds with varying species composition across the different cultivation phases of paddy. This study, conducted in the paddy fields of Kadhiramangalam, Tamil Nadu, India, recorded the bird community composition there during the various cultivation phases of paddy. The bird community data was analysed and a total of 87 bird species were recorded from the study area belonging to 41 families and 13 orders. The growth phase (PS 3) is the most diverse phase. The bird composition showed a significant variation across the paddy cultivation phases with overall average dissimilarity of 71.41%. The patterns shown by graphs of bird species composition across the paddy cultivation phases is based on guild, habitat usage and order overlap and elucidates that the change in bird community composition temporally can be attributed to the niche variability across the paddy cultivation phases. The major species contributing to these changes observed are Black-headed Munia, Baya Weaver, Common Sandpiper, Barn Swallow, Common Myna, and Black Drongo in this region.


INTRODUCTION
Birds are known to play a dual role as pests and as bio-controllers of pests in various agro-ecosystems (Borad et al. 2000). But, for decades the focus on birds in agro-ecosystems has been to study their foraging effects on crop yield and their control (Beri et al. 1968;Jotwani et al. 1969;Chahal et al. 1973;Jain & Prakash 1974;Bhatnagar 1976;Dhindsa & Toor 1980;Dhindsa et al. 1984;Parashaya et al. 1986;Subramanya 1987;Saini & Toor 1991). A few studies exist on the beneficial role of birds in agro-ecosystems (Chakravarthy 1988;Parashaya et al. 1994;Asokan & Ali 2010). The attitude on wildlife conservation became inclusive of large man-managed ecosystems (Bambaradeniya et al. 2004;Edirisinghe & Bambaradeniya 2006). Since then, the biodiversity associated with paddy fields is being considered in the light of conservation (Bambaradeniya et al. 1998;Edirisinghe & Bambaradeniya 2008;Elphick et al. 2010). Many studies on the bird use of paddy fields with focus on wetland species have been undertaken in the last two decades worldwide (Elphick et al. 2010;Sicemore & Maine 2012;Nam et al. 2015;Marco-Mendez et al. 2015).
India, being an agrarian economy, produces 21.2% of world's paddy in an area of 3.85 million hectares (Agristat 2016), making it the second largest producer of rice in the world. This large area under paddy cultivation throughout India is known to support 351 bird species (Gopisundar & Subramanya 2010). The bird species using the paddy fields are seen to vary regionally. Paddy fields are dynamic habitats and go through different habitats in a single crop cycle. This temporal variation in biodiversity during a paddy crop cycle is successive (Bambaradeniya et al. 2004). The habitat variations also lead to changes in resource availability for birds. This will have an impact on the bird community composition. As such, the bimodality in the activity pattern of birds in paddy fields during a day is known (Sridhara et al. 1983). In studying the ecological importance of birds in paddy field ecosystems, the understanding of this temporal variability in bird community would be useful. This paper aims to discern the patterns of temporal variation of bird community composition in paddy fields and explores the probable causes for the patterns observed.

Study Area
This study was conducted in Kadhiramangalam Village, Thiruvidaimarudur Taluk, Thanjavur District, Tamil . Tamil Nadu is one of the top five rice producing states in India with 2.04 million hectares (4.7% of India's paddy cultivar land) under paddy cultivation, producing 7.65% of India's rice (Agristat 2016). In Kadhiramangalam, the whole of the low lying plains are intensive agricultural areas with the major crop being paddy interspersed with very small patches of sugarcane and timber wood. The main source of water for these paddy fields is from bore wells although it is a part of the fertile Cauvery delta. Farmers used to harvest three crops in the past. In recent years, they harvest only a single crop due to unavailability of water. The fields are flooded before land preparation and later irrigated as required. Chemical fertilizers and urea are used in 80% of the fields. Pesticides are used at the farmer's discretion.

Field Methodology
To understand the bird species composition, striptransect method (Sutherland 2000) was used. Two study sites (A and B) (Images 1-4), that were more than 2-km apart, were selected in the study area. A transect of 1-km was marked in each study site. Bird data was collected for two cropping seasons of paddy cultivated from August 2016 to January 2017 and September 2017 to March 2018 from both sites. Data collected included the bird species, numbers encountered and the field variables such as field conditions (wetland, wet and dry land) and also the paddy cultivation phases.
The data has been compartmentalised into seven phases of paddy cultivation to quantify the variations in bird composition over time (Bambaradeniya et al. 1998;Paliwal & Bhandarkar 2014

Analytical Methodology
Data compiled, tabulated and subjected to basic descriptive statistics for studying the community characteristics. Pair-wise ANOSIM (Analysis of similarity) (Clarke & Green 1988) with Bray-Curtis index was used to test the significance and understand the extent of variation in the bird species composition between the paddy cultivation phases. To explore the specieswise contribution to dissimilarity, SIMPER (Similarity percentage) was used. Richness and diversity indices (Magurran 1988; Morris et al. 2014) were used to understand the temporal variation in the diversity. All these analyses were performed with PAST 3.1 (Hammer et al. 2001). The patterns in temporal variations in bird species composition, feeding guild composition and habitat usage were analysed by constructing relative abundance graphs using MS Excel 2007.

a) Bird Community Composition and Diversity
Eighty-seven bird species belonging to 13 orders and 41 families were recorded from the study area (Figs. 1a & b). Overall data shows that the passerines were the most abundant birds both in terms of species and population abundance. All species are in the Least Concern category of the IUCN Red List except Blackheaded Ibis Threskiornis melanocephalus and Rednecked Falcon Falco chicquera that are in the Near Threatened category. The basic descriptive statistics of the data compiled are summarized in Tables 1 & 2. The maximum variance and standard deviation is observed in PS 5.
The change in relative abundance of the birds as per their taxonomic order (Fig. 2a), broad feeding guild (         Diversity and Species Richness indices (Table 3) show that PS 3 (growth phase) is the most diverse with 60 species although PS 4 (flowering phase) has highest species richness and PS 2 (transplanted paddy phase) seems to be the most even. These indices also show that PS 5 (milking phase) is the least diverse with low evenness and high dominance.
The R value of ANOSIM (at 95% confidence) shows that there is a significant difference in the bird species composition between the seven phases of paddy cultivation cycle (Table 4). The average dissimilarity among the seven phases was 71.41% (SIMPER). The R values between two consecutive stages were significant except PS 5 and PS 6 ranging from 0.16 to 0.21. Between two non-consecutive stages the values ranged from 0.21 to 0.71.

b) Feeding guilds and the temporal variation
The birds were categorised into eight broad feeding guilds based on their feeding preferences in Ali & Ripley (1978) (Figs. 1c & d), viz. insectivores, granivores, carnivores, nectarivores, omnivores, aquatic carnivores (species that feed on aquatic vertebrates and invertebrates), frugivores + insectivores, and aquatic carnivores + insectivores. Considering species richness as the factor, insectivorous guild dominates (32%) as in any terrestrial habitat. Dominance of the gregarious granivorous birds is evident in the abundance pattern showing 41% of total encounters of the granivores.
The relative abundance of these guilds (based on encounter rate) across the paddy growth phases showed a four times increase in granivores from PS 3 to PS 4 (Fig.  3a). More than 50% of the omnivores declined from PS 3 to PS 4. Carnivores also declined from PS 2 onwards. The frugivores are negligible in paddy field ecosystem. The insectivores and aquatic carnivores+insectivores were observed to increase in PS 3, decrease in PS 4 and PS 5 (40% decrease) and again increase in PS 6, probably an artefact of this miscellaneous classification.

c) Wetland birds in paddy fields
Bird community of paddy fields were analysed as per their known habitat association. The 87 bird species recorded from the study area were classified into three categories, viz., Waterbirds (wetland birds), wetland-dependent birds, and terrestrial birds; and their response to the changes in paddy stages was analysed. Of these, 28 bird species (relative abundance -20%) are wetland associated, belonging to seven orders and 13 families. Twenty of these 28 bird species are true waterbirds belonging to three orders, viz., Charadriiformes -6 species (5 families), Gruiiformes -2 species (1 family), Pelecaniformes -12 species (4 families). Eight species are wetland dependent belonging to 6 orders, viz., Charadriiformes -1 species (1 family), Pelecaniformes -1 species (1 family), Coraciiformes -3 species (1 family), Accipitriformes -1 species (1 family), Gruiformes-1 species (1 family). The rest are terrestrial (Figs. 1 e & f). The relative abundance of these birds across paddy stages shows more than 80% decrease in water-birds and wetland dependent species from PS 2 to PS 6 with a

Figure 2. Temporal variation in taxonomic composition of bird community of paddy fields based on order and; a-relative abundance, and b-species richness. a b
50% drop between PS 3 and PS 4 (Fig. 4). Twenty-two species are migrants (25.2%) of which 12 species (54.5%) are wetland dependent. Nineteen species are partial migrants (21.8%) of which 10 species (52.6%) are wetland dependent.

DISCUSSION
According to Subramanya (1987), the bird community in paddy fields are bimodal across paddy cultivation phases with peaks during the tilling/levelling phase and growth phase of paddy. This pattern was observed by considering only the species richness in each of the stages. Along with the species richness the number of birds in each of the species (population abundance) is also a significant factor to explore and understand the bird life of paddy fields. Since availability of prey   is known to affect bird abundance in paddy fields (Bambaradeniya et al. 1998), it is the feeding guilds and the opportunity provided by the changing ecosystem as a substratum for feeding in the paddy fields that determine the life of birds in this ecosystem. Hence, for the better understanding of temporal variation and its significance, the abundance of each species is important along with the species richness in the paddy fields. The number of passerines increased across the cultivation phases from PS 1 till PS 5 and reduced in PS 6 and PS 7. Simultaneously, birds belonging to Charadriiformes, Pelecaniformes and Coraciiformes decreased from PS 1 through PS 5 and recovered slightly from PS 6 to PS 7. Columbiiformes showed a fourfold increase from PS 5 to PS 6 and Psittaciformes also showed a threefold increase from PS 5 to PS 7 (Fig. 2a). These results coincide with the trends observed in the guild composition variations where aquatic carnivores and insectivores + aquatic carnivores decreased through PS 2 to PS 5 with peak in PS 2. The same trends can be visualised in the wetland and wetland dependent species from PS 1 through PS 7 (Fig. 4a). The granivores showed a drastic increase from PS 3 with a peak in PS 5 and decreased in PS 6 and 7. The insectivores maintained a minimal of 15% across all the stages although the number increases which denotes their rise in abundance also across PS 1 and PS 7 (Fig. 3a).
Thus, the current study shows that there is a linear (table 3) significant change in bird community composition temporally in paddy fields along with the changes in paddy phases. This change is gradual. The richness (Table 2) did not show significant variation between the seven paddy cultivation phases considered here. So, during a cropping cycle of paddy a variety of niches are available that are also dynamic in nature. Hence, the temporal variation in bird community is due to niche variability across the different paddy cultivation phases.
The differences in bird community observed between two consecutive phases among PS 1-PS 2 and PS 3-PS 4 with R values at 0.178-0.21 (Table 3)  Increase in granivores till PS 5 and decrease only 50% till PS 7 seems to coincide with the increase in Columbidae and Psittaculidae that are seen to flock to feed on fallen grains after harvest.
The best examples of the dependency on the availability and accessibility of niches can be seen in PS 5 (milking phase) and PS 3 (growth phase). The high dominance Index value in the milking phase of paddy can be attributed to the increase in relative abundance of Passeriformes especially granivores and decrease of aquatic carnivores + insectivores (Figs. 2a & 3a). The low evenness may also be because of drastic increase in two species-Lonchura malacca and Ploceus philippinus. The steep decline in omnivores may be due to loss of open wetland conditions (Nam et al. 2015) and the crop density hinders the activities of raptors like Milvus migrans and Haliastur indus. Insectivores and mixed feeders maintain 20% of the overall abundance across the stages although there is an increase in total encounters. This shows there is an increase in the abundance of insectivores and aquatic carnivores + insectivores along the paddy stages which follow the arthropod abundance in rice fields (Bambaradeniya 1998) and changes with the habitat variations (Fig. 3a).
It can be concluded that there is a significant change in the paddy field bird composition temporarily with peak diversity during the plant growth phase (PS 3) of paddy cultivation phases. This change in bird community composition can be attributed to the dynamic habitat variability happening during paddy cultivation. Twentynine bird species contribute to 90% of the bird community changes seen in Kadhiramangalam region. The major contributing species are Black-headed Munia, Baya Weaver, Common Sandpiper, Barn Swallow, Common Myna, and Black Drongo in this region. All these except Black Drongos are colonial/ flocking birds. Hence, their presence or absence gives the major contributions. The temporal variability in the microhabitats of the paddy fields provide varied substratum in support of various bird species of different feeding guilds. This makes paddy fields a good candidate to be considered as a 'keystone habitat' for bird communities.