Lowland forest butterflies of the Sankosh River catchment , Bhutan

Acknowledgements: This study was part of the EIA of the proposed Indo-Bhutan Sankosh Hydro Electric Power project at Kerabari, being carried out by the Indian Council of Forestry Research and Education (ICFRE), Dehradun. The author is thankful to DG (ICFRE), DDG (Extension, ICFRE) & ADG and Sudhir Kumar, project manager (EIA Division, ICFRE), and Director, FRI, India, and Department of Energy and Range Officer, Kalikhola, Forest Department of the Royal Government of Bhutan for providing the necessary facilities to carry out this study.


INTRODUCTION
Reducing the impact of large dams on biodiversity calls for knowing where species, ecosystems and ecological functions are located.Lack of that information is one of the serious impediments and is a result of poor support for biological surveys, research and academic work on taxonomy and ecology.IUCN, UNEP and WCD recommendations on dams and biodiversity (McAllister et al. 2001), highlight the need to avoid biodiversity 'hotspots' and 'hot' portions of gradients.The avoidance of areas rich in species and endemic species needs to be given high priority in selection criteria.That includes both, choosing watersheds and sites within watersheds, and taking into account 'hotspots' and gradients in biodiversity.Besides, the recommendations also highlight the need to carry out research on biodiversity as our knowledge on biodiversity is incomplete and geographic distributions are poorly documented (McAllister et al. 2001).The environmental impact of large dams on lower groups of organisms, such as insects, is hardly ever studied for loss of wildlife in India (Mishra 2009).
According to Conservation International, Bhutan forms a part of the 'Eastern Himalaya Biodiversity Hotspot' which is one of the 34 biodiversity hotspots of the world today (www.biodiversityhotspots.org/).A home for over 50 species of rhododendrons, large numbers of orchids (http://www.bhutan2008.bt/en/node/276), Bhutan is also one of the 221 global endemic bird areas with over 770 avian species (Inskipp et al. 1999).Bhutan along with northeastern India are the meeting place of the central Asian and Chinese subdivisions of the Palaeartic region with the peninsular Indian and Malayan subdivisions of the oriental region and are therefore considered very rich areas in terms of butterflies.With incredible variations, from subtropical to alpine along the altitudinal gradient (100-4,200 m), Bhutan presents a large variety of habitats for butterflies.The adjoining state of Sikkim which lies west of Bhutan in India has over 689 species (Haribal 1992).Besides, 962 taxa of butterflies have been identified from northeastern India from Sikkim, Assam, Manipur; Meghalaya, Nagaland Mizoram and northern Myanmar (Evans 1932).On the other hand Wynter-Byth (1957) gives a figure of 835 species of butterflies from northeastern India including Sikkim, Bhutan and Assam up to Chittagong.However, information on butterflies of Bhutan as such is scanty.A booklet published by van der Poel & Wangchuk (2007) on butterflies of the Bhutan mountains, hills and valleys between 800-3,000 m lists only 136 species from the country.van der Poel & Wangchuk (2007) does not include butterflies from subtropical low land forests of Bhutan lying below 300m, which are contiguous with forests in India and are considered rich in terms of biodiversity.One estimate for Bhutan puts the figure between 800-900 species with the largest number being found in the 'subtropical' zone (van der Poel & Wangchuk 2007).
The present work on butterflies is part of the environment impact assessment studies undertaken on biodiversity of the influence (downstream) and impact zone (upstream) for a proposed dam to be built by India, near Kerabari (26 0 44'93N & 89 0 55'55E), in Bhutan, under the Sankosh Hydro-electric Power Project (Images 1 & 2).

Study Area
The study area falls in the subtropical forests in the southwesternmost part of Bhutan in the southcentral Dzongkhag (Sarpang District).These forests are broadly classified as semievergreen but vary from almost totally deciduous on exposed dry slopes to almost totally evergreen in the forest valleys.Adjacent to and east of the study area lies the Phibsoo Wildlife Sanctuary.The sanctuary covers an area of 265km 2 of unique dry shorea forest ecosystem and conserves tropical fauna such as the Tiger Panthera tigris, Elephant Elephas maximus, Gaur Bos gaurus, Golden Langur Trachypithecus geei, Indian Wild Dog or Dhole Cuon alpinus, Grey Peacock Pheasant Polyplectron bicalcaratum, Rufous-necked Hornbill Aceros nipalensis, Wreathed Hornbill A. undulatus, Great Hornbill Buceros bicornis and is the only protected area in Bhutan having Chital Deer Axis axis and natural sal forests.Phibsoo Wildlife Sanctuary touches its border with India and is linked on its southwestern end to the Buxa Tiger Reserve in West Bengal, India ca.100km to the south-east of the study site, in India, lies the Ripu-Chirang Wildlife Sanctuary (RCWF; 89 0 55'-90 0 30'E & 27 0 15'-26 0 35'N) in western Assam which is a transitional zone between Manas Tiger Reserve in the east and Buxa Tiger Reserve in the west.About 300 species of butterflies have been recorded from RCWF including the rare Yellow-crested Spangle Papilio elephenor Doubleday, 1886 and Moore's Cupid Shijimia moorei Leech, 1889 (Choudhury 2010), although the list has not been given.

Sampling
Five sampling surveys were carried out during 7-9 January 2009 (winter), 13-14;17-19 May 2009(pre-monsoon), 11-13 July 2009(monsoon), 26-28 November 2009(post-monsoon) and 19-22 March 2010 (spring) in the study area .In January the day temperature and relative humidity (11:30-15:30 hr)   with high relative humidity (83-90 %) and temperature (30.3-32.1 0 C).At this time rain drizzled intermittently throughout the day while all the road network between Kalikhola (26 0 44'20N & 89 0 51'25E) and Kerabari was washed out at places by running seasonal streams.During November the weather was cool and dry (relative humidity 61-62 %; temperature 26-27 0 C) the weather at this time being more comfortable to work in and the road network is also restored.Transect walks were undertaken along the road that links Kalikhola Village located on the Indian border to Kerabari Village (15km) and then further on a foot trail linking Kerabari-Huma Semal Village (6km) along the Sankosh River and finally 5km further north upstream from Huma Khola on a foot trail.All the sampling sites lay on the right bank of the river Sankosh between 118-220 m and 26 0 44'21"-26 0 47'69"N & 89 0 51'25-89 0 56'05"E.'Pollard Walk' method used for sampling butterflies was carried out throughout the day from morning (10:00 hr) until sunset (17:00-18:00 hr), but the total number of sampling hours varied from 4-8 hr per day being less during the monsoon season (July=3.5-4h/day)and maximum in the premonsoon (May=7-8h/day).Thus, a total of ca.110 hr of sampling was carried out during the entire study period.Butterflies were recorded up to 20m on both sides of the transect and on the river bed of Huma Khola near Huma Village.In addition, a transect (500m) uphill along the forest stream which flows into the river Sankosh at the U-turn near Kerabari was also sampled by trekking for a day (04 March 2011).During each season sampling was carried out for 3-4 successive days at a stretch on these trails.Butterflies were identified mainly by taking photographs, and using field guides for identification (Evans 1932;Wynter-Blyth 1957;D'Abrera 1982D'Abrera , 1985D'Abrera , 1986;;Smith 1989Smith , 2006;;Haribal 1992;Kehimkar 2008).Identification of some species of the Neptis soma group (Nymphalidae) and the Telicota colon group (Hesperiidae) was based on wing pattern as captured in the images, and not based on the genitalia. Hene these might require confirmation.

Data Analysis Species Accumulation Curve
Data for the number of species recoded in each season was pooled.Species accumulation curve was then plotted seasonally from the first to the last sampling to see the rate of species accumulation during each of the five successive samplings.

Relative Abundance
The abundance data for each butterfly species for all the five seasons combined was pooled separately to get the overall relative abundance status of each species.This data was then sorted in descending order from highest to lowest value and species were then placed in four different classes based upon their respective quartile divisions from very common to uncommon, respectively i.e.VC -very common or Q1 (7-217 number of individuals sampled); C -common or Q2 (4-6); FC -fairly common or Q3 (2-3); UCuncommon or Q4 (1), respectively.

Student's t-test
Relative abundance data of 'very common' species (Q4; n=48) of butterflies for individual seasons (n=5) was then subjected to 'paired t-tests' (both twotailed and one-tailed) to see if the seasonal variations between two different seasons were significant at p< 0.05 using SPSS 14 software.[Data was tested for normality prior to analysis and non-parametric tests were followed accordingly to look at the variations].

Simpson's Index
As species richness and evenness increase, diversity increases.'Simpson's Diversity Index' takes into account both richness and evenness (Evenness is a measure of the relative abundance of the different species making up the richness of an area).'Simpson's Index of Diversity' was calculated and used here to see the seasonal variation /trend in species diversity of butterflies in the lowland forests of Bhutan.Simpson's Index (D) (Simpson 1949) measures the probability that two individuals randomly selected from a sample will belong to the same species (or some category other than species).D = ∑ n(n-1)/N(N -1) n = the total number of organisms of a particular species N = the total number of organisms of all species The value of D ranges between 0 and 1.With this index, 0 represents infinite diversity and 1 no diversity.That is, the bigger the value of D, the lower the diversity.As this is neither intuitive nor logical, to get over this problem, D was subtracted from 1 to give Simpson's Index of Diversity = 1-D.The value of this index also ranges between 0 and 1, the greater the value, the greater the sample diversity.In this case, the index represents the probability that two individuals randomly selected from a sample will belong to different species.

Sørensen's Similarity Index
This index or ß was calculated to see the species similarity of butterflies between different seasons in lowland forests.ß = 2c / (S1 + S2) Here, S1= the total number of species recorded in one season, S2= the total number of species recorded in a different season, and c=the number of species common to both the seasons.
The Sørensen's similarity index (Sørensen 1948) is a very simple measure of beta diversity, ranging from a value of 0 where there is no species overlap between the communities, to a value of 1 when exactly the same species are found in both communities.

RESULTS AND DISCUSSION
A total of 213 species of butterflies were recorded during the five sampling surveys carried out in 18 days in the study area.A complete checklist of butterflies sampled in the study area is given in the Table 1).Amongst these, only 87 species are common with van der Poel & Wangchuk (2007) list for areas between 800-3,000 m in Bhutan.

Species accumulation curve
The increasing trend in the species accumulation curve shows that new species were added during every season up to the last sampling at about the same rate (Fig. 1).This suggests that there is a potential of adding more species to the current list of the area and there is a need to undertake more surveys for a preparing an exhaustive list.

How many species could be there in the study area?
A total of 22 species of the family Papilionidae were recorded in this which suggests that the species richness of the area could be as high as 297 species based on family proportion model (Singh & Pandey 2004) by taking Paplionidae species proportion as 7.4% of the total for northeastern India (Wynter-Blyth 1957).The present findings thus represent only 70.6% of the species found in the study area.The family Pieridae represents only 10.8% of the total; Lycaenidae 24.0%; Nymphalidae 42.4% and Hesperidae 13.3%, respectively, of the total species sampled.As per Singh & Pandey (2004) model, families Lycaenidae and Hesperiidae should represent 29.5% and 21.9%, respectively, of the total number of species sampled in northeastern India.These two families are thus under represented in the present sampling and there is a need to look for more species.

Seasonality of butterflies in subtropical lowland forests
Species richness and diversity: The seasonal variation in Simpson's Index of Diversity (Fig. 2) and the number of species sampled during each season (pooled data; Fig. 3), suggests that species diversity and richness both peak during March (spring) and are lowest in January (winter).A second peak was observed during November which was smaller than the first peak.This seasonal pattern of diversity in butterflies is very typical of the lower foothills and valleys of the Himalaya, i.e.Dehradun Valley (400-700 m) (Singh & Bhandari 2003) or even Calcutta (Wynter-Blyth 1957).However, the timing of the two seasonal peaks, as observed in this study, is slightly different in pattern as compared to the butterflies found in the higher reaches of the Himalaya where the first peak is slightly late in April-May while the second peak is slightly earlier in September-October i.e.Bhutan between 1,200-3,000 m (van der Poel & Wangchuk 2007) or Shimla at around 2000m (Wynter- Cumulative number of species Polled data for each season Blyth 1957).This is because spring arrives late in the higher reaches so butterflies emerge later and as winter sets in early in the hills, so butterflies go into hibernation early as compared to the low land forests.
Species similarity between seasons: Sørensen's similarity index between seasons varied between 0.3076 to 0.5656.This suggests that the species composition also changed with the seasons in these forests.Greatest species similarity was observed during January and November followed by March and November, respectively (Fig. 4).However, the least species similarity in species composition was observed between January and July followed by January and May, respectively (Fig. 4).
Relative abundance: The five most abundant species in the study area were Euploea mulciber mulciber, Ixais pyrene pirenassa, Appias lyncida hippoides, Tirumala septentrionis and Eurema blanda silhetana.The seasonal variations in relative abundance of butterflies for 'very common' species (n=48; Q4) were found to be significant (p< 0.05; Student's t-test) between only January-March; March-May and March-November.Thus, the abundance of common butterflies during 'spring' varied significantly as compared to 'winter', 'dry summer/pre-monsoon' and 'post-monsoon' seasons.
Species of conservation priority: At least 11 species (Pareronia avatar avatar, Nacaduba pactolus continentalis, Porostas aluta coelestis, Lampides boeticus, Melanitis zitenius zitenius, Elymnias vasudeva vasudeva, Mycalesis mestra retus, Charaxes marmax, Athyma ranga ranga, Neptis manasa manasa and Neptis soma) recorded in the study area have been listed in Schedules I and II of the Indian Wildlife (Protection) Act, 1972 (IWPA), but none in the 'Forest Nature and Conservation Act of Bhutan 1995'or IUCN Red list of Threatened species (IUCN 2010).Amongst these 11 species, except for Lampides boeticus , all other taxa are 'rare' throughout their distribution range (Evans 1932;Wynter-Blyth 1957) and are thus species of conservation priority over rest of the other taxa in the study area.Besides, one more species, Burala amara Moore, which has been omitted from the IWPA, is also  a 'rare' species in India (Evans 1932).

CONCLUSION
High value of 'Simpson's index ' (0.8929-0.9687) points to the area having high variation.A stretch of ca.12km monitored during the 18-day sampling revealed as many as 213 species and had the potential of adding many more species.The list provided here is in no way complete but represents ca.2/3 species found in these tracts in the study area and ca.¼ of the species estimated from Bhutan.Besides, 12 taxa that are 'rare' in occurrence throughout their distribution range and worthy of conservation were also recorded in this subtropical lowland forest tract.Also, a good number of butterfly species were present during all the five seasons (93±12 species per season; range = 66-129 species) which is a good indicator of the potential of this area for butterfly ecotourism, that needs to be explored.'Biodiversity offsets' such as butterfly conservatories or even a butterfly park can be established here for conserving the 'rare' species and promoting ecotourism.Thus, the present study is a way forward to bridge the remaining gaps today in documenting the complete butterfly fauna of the area, identifying sites and species of conservation concerns in the subtropical lowland forests of Eastern Himalaya.
of Bhutan depicting the location of protected areas and the study area bordering India Image 2. Location of study sites, the road from Kalikhola to Kerabari and the foot trail beyond up to Huma Village, marked in blue, along which the sampling surveys were undertaken along the right bank of Sankosh River.

Figure 1 .
Figure 1.Species accumulation curve in sampling butterflies through different seasons in tropical lowland forests of Bhutan (January 2009-March 2010).
Figure 2. Seasonal variation in (A) species diversity of butterflies, and (B) its standard error in subtropical lowland forests of Bhutan (January 2009-March 2010).